Principles and Rules of Research Methodology - Essay Example

?Discuss the advantages and disadvantages of one or two particular research methods (e.g. ethnography; narrative enquiry; surveys; questionnaires; interviews). Refer to three or four relevant research articles in an area you may be interested in investigating and consider the strengths and weaknesses of the methods used in each context. Survey Research The survey research methodology mostly started with social science studies where it was applied for research on businesses, marketing trends, politics, and economics (Marelli, 2004). As these areas evaluate human behavior with most public actions founded on behavior, social science principles on survey research are very much applicable to public health. Social science survey research is significant as it seeks to measure previous and current behavior, as well as future behavior (Marelli, 2004). As can be seen from political survey polls, these polls evaluate public opinion likely determining how people would vote in upcoming elections. The surveys also in public health would assess health behavior including eating trends, exercise trends, smoking behavior, wearing of seatbelts, frequency of routine exams, and other related relevant health behavior (Marelli, 2004). These studies seek predictions and trends in human behavior and how such behavior can be changed. In effect, individuals without health insurance can be surveyed using questions on how willing they are to pay health insurance. The surveys can also be applied in determining if individuals would buy more fresh vegetables if these were cheaper (Hatch, 2009). Survey subjects likely covered for survey research include those who can be accurately evaluated through subject response (Hoppe, 1998). Physical measures can sometimes be accurately assessed if the subjects have standard resources. In other words, height, weight, blood sugar, nutrition, and blood pressure can easily be measured where the subjects apply standard measures. Medication use can also be evaluated based on an individual’s prescription information. Clinical results, however, are sometimes not being measured during the survey research especially as they need health care professionals to carry out the process (Mathiyazhagan and Nandan, 2010). Health care application would fit well for surveys since the responses can be taken from the subjects, insurance records, or the written records. Disease incidence and prevalence are often more complicated to measure as they call for accurate means of diagnosing (Hatch, 2009). Mental health diseases are considered exempted because they are often diagnosed using survey scored within or outside the mental health system. Validity The validity of survey research is evaluated by measures similar to clinical research studies (Hatch, 2009). The covered population must still be specifically defined and a sample population extracted. The most relevant source of bias in survey research is on sampling. As soon as the target population is established, random or representative samples have to be chosen and those chosen must then be asked to participate in the study (Glasow, 2005). For some populations, the samples are easy to define and extract. Samples in clinical settings can be randomly chosen from the medical records. Inclusion and exclusion criteria, including condition, age, gender, treatment can then be compiled before the sampling process is initiated (Hatch, 2009). Patients with existing relations with the researcher would likely entertain calls or invitations to participate. The larger population however would be more difficult to locate and include in the sample population. Granting that the target is a population of adults in a city in the eastern coast of the United States, the first task would be to identify all the adults in the area. The sources would also have to be laid out. Telephone listings are considered one of the sources of names, however not all adults have listed numbers, and others use cell phones, and the rest, do not even have phones (Hatch, 2009). In some instances, other databases may be considered. For individuals with businesses, they would likely have listed numbers and would have a greater chance of being included in the population. Census records can also be used as an alternative to sample selection, however, these records may not be updated (Hatch, 2009). Moreover, transient and homeless individuals would likely not be included in the census records as they do not have a permanent address. The manner of contacting the subjects can also cause sample errors to emerge. In cases where telephone surveys or door-to-door personal surveys are carried out, the day of the week and time would also have to be determined (Tourangeau, 1999). Surveys carried out during the weekdays would likely not include working adults; with most of the covered adults stay-at-home parents or elderly adults. Surveys carried out during the night hours would likely lead to shift-work adults being covered in the sample population (Hatch, 2009). Adults who stay at home and who may be retired have different experiences as compared to the working adults, and vice versa. As such, the results for any survey conducted on these groups of individuals would likely yield different results. The most significant challenge in the conduct of survey research relates to response rates. Possible respondents may sometimes be unreceptive in answering the questions (Converse and Presser, 1986). In some cases, getting unidentified callers would likely prevent the potential respondents from answering the calls. Mail surveys are usually ignored. In cases where persistent and multiple attempts are made by the researchers, the response rates may manifest improvements; however, majority of these surveys do not often get 100% response rates (Hatch, 2009). This would then raise the issue that those who may be unwilling to join the surveys may not have the same responses as the ones who are willing to participate. The possible disparity in responses would then possibly cause sampling errors in the study. Increasing response rates by offering incentives or reminding respondents have also been effective. Mail surveys often have a 10% respondent rate initially, however a follow-up letter which would offer incentives or phone call reminders have been known to increase response rates (Kalton, 1983). The incentives can take on many forms including cash, gift certificates, or shopping discounts. Other surveys also promise the issuance of checks or other cash incentives once the survey is finished. Another problem involving the use of the survey method is the misclassification or information bias. This usually relates to errors in the evaluation of variables (Attewell and Rule, 1991). In instances where the respondents are measured in terms of demographic qualities, their weight, height, blood pressure, or temperature, the results may not always be consistent especially as different scales of measurement may be applied (Hatch, 2009). In cases where the results would take different directions, there is a random misclassification; this may not impact on the results of the study especially when the sample size is large. However, where the standards of measure generally and consistently over or under estimate within the measured population, the misclassification would actually impact on the results (Hatch, 2009). For instance, where glucometers would under estimate blood sugar for individuals without diabetes and over estimate it for those with diabetes, then there would be a higher gap in the statistical measure or error (Hatch, 2009). Recall errors are also a huge issue for surveys. Human memory cannot recall all details of one’s life, especially those which are not important (Browne and Keeley, 1998). The more recent the event, the higher would be the recall possibility. In cases where simultaneous incidents occur, memory tags are usually essential in highlighting the memory which needs to be recalled. Most events however unfold daily and most surveys seeking information from previous behaviors would likely be affected by inaccuracies (Fowler and Floyd, 1995). The questionnaires used for the surveys as well as the methods of information extraction would likely cause misclassification. Questions in surveys are called items and establishing a single interpretation for each item is a difficult process. In the question raised by a study on optometry (Murphy and Addis, 2008), the question asked was, “Is private practice superior to corporate practice?” This is a subjective question and would likely be interpreted in different ways. The question of what ‘superior’ actually means is one of the issues seen in the question. It may refer to a better quality of life, a better regard from colleagues, or even higher income, or better personal satisfaction. The elements of private and corporate practice would also have to be determined. Optometrists are usually engaged in private practice in retail institutions and most of them would define this to be private practice; but others would define such as corporate practice (Murphy and Addis, 2008). The biggest issue in the survey is on the sampling. The questions were printed in the journal and the subject-respondents were selected. This implies that the opinions drawn were founded on a population whose motivations may be different from the motivations of optometrists in general. In effect, the results cannot therefore be applied to other populations, other than the current optometric respondents. Writing items for questionnaires can be significantly challenging. The answers chosen may be dichotomous, may refer to yes or no responses, high or low, present or absent, or in other cases, never sometimes, frequently, and always (Hatch, 2009). A choice may also be forced, with the respondent being asked to favor one side over the other, especially with choices like strongly agree, agree, disagree, or strongly disagree. In essence, the short questions focused on narrow subjects are most appropriate. Answers may also overlap, especially in instances where the subject’s age is being asked and the choices include a range of years. In some instances, the grouping of age may overlap, where a 10 year old may be categorized in the 0-10 years or the 10 to 20 year category. A correction of entry may help resolve this matter however, there may still be an error, especially in cases where the respondents nearing their birthday may already choose the next higher range. It may also be significant to include an introduction to the survey, mostly as a means of helping the respondent understand the topic of the survey and its goals (Converse and Presser, 1986). Such introduction may also be made on each section of the questionnaire, especially where shifts in topic are seen. Despite these specific elements included in the survey process, respondents would still likely seek clarification. It is crucial therefore, especially if surveys are carried out orally, not to make an interpretation of the question for the respondent (Kraemer, 1991). Where the subject would try to seek clarification, the researcher must only restate the questions as asked. This would avoid any undue influence from being seen in the questions raised (Kraemer, 1991). To prevent vagueness in the process, items or questions raised would have to be clear of double negatives, implied negatives, multiple meanings, and knee-jerk options. (Converse and Presser, 1986). Double vision must also be avoided, especially as the respondent would not know how to answer the single item which raises two or more different issues. The words “or” and “and”, including “either” would likely present double vision for the respondent (Converse and Presser, 1986). Separating the items for the questionnaire would easily resolve this issue. In cases of implied negatives, there may be a suggestion of one being better than the other. The subject may answer in the way which may be suggested strongly. The question: Do you wear your contact lenses longer that the doctor-recommended period? is a badly phrased question as there is already a suggested answer or direction to the question raised (Hatch, 2009). The better phrasing would be to ask: “In general, how long do you wear your contact lenses?” and choices with time range can be placed after the question. Leading questions are all too common in surveys and for the researchers, it is important for them to avoid these questions in order to ensure that the answers would be valid (Hatch, 2009). A Veteran’s Affairs low vision clinic commissioning a survey set out to evaluate the impact of low vision rehabilitation for patients (Hoppe, 1998). The survey sought to determine how patients utilize the devices and whether their behavior was improved. The sample was randomly selected from chart numbers of patients in the clinic. All patients were surveyed. Most of the subjects were retired and a phone survey was chosen as the data gathering tool. In general, the questions and formats were developed for the survey (Hoppe, 1998). The above process indicates a means by which patients can be measured in terms of their use of the devices and if there were eventually changes in the devices seen. The public health impact of the study includes the assessment of the patient quality, assessment of cost benefit and the evaluation of independence effect (Hoppe, 1998). In the current era of evidence-based practice, more researches have been deemed necessary in justifying services to patients as well as third parties. Advantages for survey methods are however also very much apparent despite the issues raised above. First of all, the survey allows for the collection of a significant amount of information, quickly and with the least cost (Binder, 1998). Information can also easily be obtained from large groups of people, especially those who may be located across a wide geographical area. The respondents may also be able to easily answer the survey within a short amount of time and at a preferred choice (Binder, 1998). These surveys also allow a more significant input and for a consensus to be established from the study results. The survey questionnaires can also be adjusted based on a specific group of respondents (Sanglimsuwan, 2011). Anonymity of answers is also easier to ensure in surveys, with candid and honest answers assured. Multiple choices or rating type questions often produce quantitative data which can easily be summarized and collated, and easily be presented in the form of tables and graphs (Sanglimsuwan, 2011). There is also better flexibility in the assessment of behaviors and the collection of information. In a survey research by Gregg and colleagues (2012), the authors set out to establish the trends in death rates among US adults with and without diabetes between 1997 and 2006. The survey was also based on the applications of the National Health Interview Survey. The NHIS is considered a regular survey on health status and health behavior for the US general population (Gregg, et.al., 2012). It applies multistage probability sampling methods in order to choose about 41,000 households annually in the US. Their annual response rates from 1987 to 2004 registered at 87 to 92%. In this diabetes status survey, the authors covered about 240,000 adults 18 years old and above randomly chosen from each of the 41,000 households chosen as respondents (Gregg, et.al., 2012). About 89% of the data from the respondents were deemed sufficient in securing accurate results (NHIS, 2009). All NHIS surveys are subjected to oversight; and participants were also asked to give their consent. The interviewers evaluated the diabetes status of the participants by asking them if a doctor of any other health professional have told them that they had diabetes or hyperglycemia (Gregg, et.al., 2012). The number of years since the diabetes diagnosis was made was also recorded. Moreover, the respondents were also interviewed about their age, gender, education, family income, history of CVD, and their self-reported height and weight. Income was used as a determinant of income or poverty status and need. A score of 1 indicated the threshold for poverty, or a relative level of poverty, and a score of 0.1 indicating income on multiple poverty levels (Gregg, et.al., 2012). Mortality was based on the National Death Index, also based on the NHIS standards. The vital status of the respondents was based on their time survey from 1997 to 2004. The main goal of their study was to determine whether the death rates in the recent samples of respondents were not the same as the rates seen in earlier samples. The authors also assessed whether the changes in the mortality rates were seen between the diabetic and nondiabetic population, on elements and qualities including age, race, gender, and socioeconomic groups affected. The researchers also grouped the survey participants into four consecutive 2-year samples with death rates measured for the population. Death rates were also based on age-standards on projected year (Gregg, et.al., 2012). ANOVA tests were also applied in order to establish whether demographic and health qualities were different across surveys. The authors also compared survival distributions between samples based on their diabetes status and covariate qualities. In critically evaluating the study above, it is important to note that even as the NHIS has the biggest and most representative data on diabetes, these findings are restricted by the self-report nature of the answers. As such, about 20% of diabetes cases are actually undiagnosed. Since the portion of diabetes incidents which are still undiagnosed would likely be increasing, more recent cohorts relating to diabetes may be affected by people diagnosed with diabetes earlier, likely leading to lower mortality rates (Isaac and Michael, 1997). Moreover, diabetic patients who were not aware of their status would not have been misclassified into the nondiabetes group. Since the group is likely to incur a higher death rate relating to individuals not having diabetes, the high mortality risk referring to diabetes would likely be an underestimate. To evaluate this, a sensitivity assessment was carried out where those who had diabetes for 2 years of less were excluded from the recent sample drawn. The results did not make much difference on the findings. This implies that the decrease in death rates within the samples do not seem to be explained by previous diagnosis. Not much difference was also noted in the prevalence of CVD or functional difficulties for the samples. Still, to eliminate bias caused by the earlier detection in the samples, studies would have to relate the death rates with the cohorts which include respondents having undiagnosed diabetes. This analysis nevertheless represented a larger cohort of individuals with diabetes, alongside a national probability sampling tool. In effect, the study would be the most representative evaluation on the death rates of the US diabetic population. In another health-related survey research, authors Liu and colleagues (2010) set out to describe the prevalence of chronic complications seen in urban Chinese type 2 diabetic outpatients, and to assess the issues between chronic complications and the patients’ demographics, including diabetic-related clinical characteristics. The study used a cross-sectional hospital-based research covering 4 major Chinese cities of Shanghai, Chengdu, Beijing, and Guangzhou (Liu, et.al., 2010). The survey covered March to July 2007, with 1524 type 2 diabetic outpatients included in the respondent population. A face-to-face interview carried out by trained interviewers was applied, mostly with a questionnaire covering data on demographics, disease presentation, as well as complications. The HbA1c test was also carried out for each of the respondents. The study later arrived at the conclusion that chronic complications are very much prevalent for type 2 diabetic outpatients with blood glucose control for these patients being very poor (Liu, et.al., 2010). The authors recommended that efforts must be directed towards ensuring intensive blood glucose control, improving early diagnosis and strengthening the case management of the disease in order to minimize if not totally eliminate the manifestation of complications. In evaluating the above study, there are immediately issues and weaknesses which are noted. Initially, as the study was hospital-based and the respondents were within the general hospital settings of major cities of China, these results may only be applicable to type-2 diabetes related chronic complications within the population and region covered. In effect, the findings would apply to hospital-managed patients, not the general population of diabetics. Moreover, as patients who do not often visit the OPD would likely be healthier than those included in the study, the study is most likely making an overestimation on the exact prevalence of chronic type 2 diabetes-related issues among the diabetic outpatient population in mainland China. Another noted issue in this study by Liu, et.al., (2010) is based on the fact that purposive sampling was used, not random sampling. Service assessment, capacity in the management of type 2 diabetes and the participation intention were evaluated during the sampling process. The chosen hospitals were mostly tertiary or secondary hospitals having specialized units for diabetes and most of the diabetic patients being treated were visited. Still, the purposive sampling may still impact on the results, mostly in terms of generalization (Glassner and Moreno, 1989). Issues in this survey research also refer to the fact that the authors did not have any data on patients who refused to participate in the research. To improve response rates, the data collectors waited for the patients outside the consultation rooms, and asked patients for interviews following their consult. Not many patients rejected participation however, bias during the recruitment process may already be incurred, especially in terms of recruitment methods. Case Study Research Case studies are generally considered qualitative studies because they focus more on the quality of the answers, not so much on establishing trends supported by statistically sound results. Case studies are considered flexible (Hsieh, 2005). As such, the researchers are able to choose a topic and indicate its boundaries. The methods are also more flexible as compared to quantitative studies, including survey researches. Numerous types of data collection processes would also be collected based on how the researcher himself perceives the best revelation for reality (Hsieh, 2005). The manner of collecting data is also highly flexible as there is no designated time and end of collecting needed information. The subject respondents can therefore be changed. The case study method can also be carried out independently, or as a part of a larger research. Case studies are also able to go in more detail during the research process. Data are gathered over a certain period and are usually retained in the words of the respondents (Bassey, 1999). These case studies are also usually carried out in order to assess a program or to evaluate a hypothesis and devise a possible concept or theory. These case studies also cover contemporary events in society. They therefore investigate incidents which occur in their natural settings (Bassey, 1999). In general, the above points are considered the strengths of case studies, however other researchers may label these as weaknesses. Case studies however also have their weaknesses. One of the primary weaknesses refers to generalization (Denscombe, 1998). Due to the limited sample size which is usually covered by case studies, the results would not be applicable to a larger population. Moreover, the subject-respondents are often not chosen randomly, but are chosen based on specific qualities being sought by the researchers (Denscombe, 1998). In effect, the results of a single case study may only be applicable to the subjects being covered. There may therefore be some doubt or disbelief in any theories or principles drawn out from these case studies (Jensen, 2001). Other researchers have sought to improve the generalizability of these case studies by using clusters of respondents and later accumulating results from these clusters for a wider and more general application. So far this practice has not gained much support from researchers, mostly because the generalizability of case studies would not be improved by the clustering of respondents (Jensen, 2001). A population to population accumulation would also not improve the reliability of the results. Reliability is also an issue in case study research. Reliability refers to consistency observed in the experimenters on different occasions and different observers (Merriam, 1988). Validity would refer to how accurately an account would highlight social phenomena. Reliability and validity are usually assessed together especially as they relate to similar issues all impacting on credibility of research (Merriam, 1988). Reliability and validity may also be regarded differently by researchers based on numerous contexts. The issue of reliability and validity in case study is based on the lack of operational standard measures. Various techniques in securing reliability and validity of case studies have been suggested, especially in terms of controlling the subjectivity of data and ensuring critical evaluation of such data gathered (Lazaraton, 1995). In general, the critical evaluation of data gathered would help contradict any inconsistencies in the data, allowing for gaps in the research to emerge and be managed. Various studies in psychology and mental health research have been carried out using case study methods. A critical analysis of the case study application will now follow. In the study by Hansson, et.al., (2011), the authors set out to define the formation and structure of coordination in a health and social care consortium for individuals with mental health problems in one area of Stockholm. This study was also meant to evaluate the effect on care and client results. A multiple-method case study design was applied with data gained through interviews and informants from various organizations (Hansson, et.al., 2011). The study indicated activities and factors which assisted and interfered with the formation of coordination activities. A major barrier was on the focus of the county of on purchase volume and costs of services, without focusing much on the quality of services. In evaluating the study, there are concerns on how the boundaries are established between the content, the process, as well as the context and outcomes of the study (Hansson, et.al., 2011). In this case, the researchers sought to understand the case based on the correlations of individuals with mental issues in the region, and to understand the other changes were caused by the other designated changes. The researchers actually sought to identify late systems changes; mostly referring to shared care plans mostly as an intermediate outcome (Silverman, 2005). In further understanding the case study, the primary goal of the case study was to define the links between the elements which would have assisted and those which would have interfered in the implementation of changes. The study design was also not extensive enough in identifying other changes which would have impacted on the service outcomes (Silverman, 2005). The case study also suffers from a limited number of respondents. A cross-section interview of the participants decreased the issue of limited sample for this case study to some extent. Further studies were understood in the analysis of this case study, mostly in terms of external validity and on determining whether the findings would be applied outside the population being studied (Silverman, 2005). It is important to understand that case studies often focus on analytical generalization, not so much on statistical generalization (Yin, 2009). To some extent, this worked well for the study especially as analytical generalization helped correlate the results to the appropriate theory. Some of these elements would translate well to the question of cause-effect relationships. In another case study, Griffiths (2009) set out to determine the role of context in differentiating between excessive gaming from addictive gaming. The gamers in this study apparently played up to or even more than 14 hours a day. They were however different in terms of motivation and meaning as well as experience of gaming (Griffiths, 2009). One of the players actually appears to be very addicted to online gaming however the other player is not that addicted to online gaming, especially when applying context and consequences. The two cases also emphasize the significance of context in the life of gamers and indicate how excessive gaming does not often imply that an individual is addicted (Griffiths, 2009). The study also argues that online game addiction must not be defined by the extent to which the excessive gaming would impact on other elements of a gamers life; instead the focus must be on the amount of time spent playing. The authors also concluded that activities cannot be considered addiction if there are no negative effects in the individual’s life, even if the player is still playing up to 14 hours daily. There are significant weaknesses noted in this case study, the main one being the limited sample population (Yin, 2009). There were only two subjects included in the study. Two research subjects would likely fulfill the goals of the research, but would not enrich the wider knowledge on the subject matter of online addiction to gaming. The analysis of the results is also based on general parameters which do not necessarily measure the holistic impact of online gaming on individuals (Yin, 2009). The self-report from the respondents also do not necessarily support accuracy in the results, especially in terms of the personal impact of online gaming to their social and family life. The relatives or the friends of the respondents may actually respond differently when queried about how the online gaming has impacted on their relationship with the subject respondent (Silverman, 2006). Moreover, 14 hours of play cuts into the respondents’ time for his social and family life. Indicating that the 14 hours of gaming has not affected the individual’s social and family life does not seem to be a realistic result. It does not provide a dynamic and contextual finding fully supported by a holistic assessment of the subject in his social and family setting (Silverman, 2006). Despite such weakness however, the strength of this study is based on the fact that the respondents were able to express in their own words the impact of online gaming in their lives. This provides a more personal application of case studies as far as specific populations are concerned. Although these results would not apply to a larger population, the information can still be used in understanding individual respondents, especially those who share similar qualities as the respondents in this case (Silverman, 2006). The study by Wild and Ehlers (2010) declared that the Cognitive Therapy for PTSD developed by Ehlers and Clark in 2000 was an effective therapeutic resource when carried out weekly and daily. It is however not determined whether patients with PTSD could benefit from a self-study assisted cognitive therapy. Their case study then set out to study this possibility in treatment (Wild and Ehlers, 2010). The case study covered one patient with PTSC and comorbid major depression who manifested these issues after a road accident. He was then managed in six sessions of cognitive therapy with six self-study modules carried out in between the sessions. The patient was able to recover completely based on measures available for PTSD and depression as assessed through self-report and independent evaluation (Wild and Ehlers, 2010). Once again, this study has an even smaller sample size, with only one patient made subject of the research. As such, the generalizability of the study is very much limited. The results cannot be applied to a larger population and can only actually be applied to the individual subject of the study (Crowe, et.al., 2011). It is important to note that case studies are however able to adapt the study process to fit the individual’s case. This cannot be carried out in quantitative studies where fixed and standard measures are applied to respondents regardless of their specific qualities. The case study was also able to illustrate the benefits of using self-study modules in the management of PTSD even with memory and concentration issues seen in the patient group (Crowe, et.al., 2011). The conclusions drawn are however very much limited and it is difficult to determine if the treatment actually can be credited for the patient’s recovery. This case study is however a guide for future studies on the subject matter, especially as gaps in the research are very much apparent and can later be used as tools for improvement in randomized controlled trials which can be carried out in the future (Crowe, et.al., 2011). References Attewell, P. and Rule, J. (1991). Survey and other methodologies applied to IT impact research: Experiences from a comparative study of business computing. Paper presented at The Information Systems Research Challenge: Survey Research Methods. Bassey, M. (1999). Case study research in the educational setting. Buckingham: Open University Press. Binder, D. (1998). Longitudinal surveys: Why are these surveys different from all other surveys? Survey Methodology, 24(2), pp. 101-108 Browne, M. and Keeley, S. (1998). Asking the right questions: A guide to critical thinking. (5th Ed.). Upper Saddle River, NJ: Prentice Hall. Converse, J. and Presser, S. (1986). Survey questions: Handcrafting the standardized questionnaire. Newbury Park, CA: Sage University Publications. Crowe, S., Cresswell, K., Robertson, A., and Huby, G. (2011). The case study approach. BMC medical research methodology 11(1), p. 100. Denscombe, M. (1998). The good research guide: For small-scale social research projects. Buckingham: Open University Press. Fowler, J., Floyd J. (1995). Improving survey questions: Design and evaluation. Thousand Oaks, CA: Sage Publications. Glasow, P. (2005). Fundamentals of survey research methodology [online]. Available at: http://www.mitre.org/work/tech_papers/tech_papers_05/05_0638/05_0638.pdf [Accessed 29 May 2013]. Glassner, B. and Moreno, J. (1989). The qualitative-quantitative distinction in the social sciences. Boston: Kluwer Academic. Gregg, E., Cheng, Y., Saydah, S., Cowie, C., et.al., (2012). Trends in death rates among us adults with and without diabetes between 1997 and 2006: Findings from the national health interview survey. Diabetes Care, 35(6), pp. 1252-1257 Griffiths, M. (2010). The role of context in online gaming excess and addiction: Some case study evidence. Int J Ment Health Addiction, 8, pp. 119–125. Hansson, J., Vretveit, J. and Brommels, M. (2012). Case study of how successful coordination was achieved between a mental health and social care service in Sweden. International Journal of Health Planning and Management, 27, e132-e145. Hatch, S. (2009). Study design for survey research: Optometric care within the public health community [online]. Available at: http://webpages.charter.net/oldpostpublishing/oldpostpublishing/Section%202,%20Principles%20of%20Public%20Health/Sect%202,%20Study%20Design%20for%20Survey%20Research%20by%20Hatch.pdf [Accessed 28 May 2013]. Hoppe, E. (1998). Survey research methods. In: Hatch SW. Ophthalmic research and epidemiology. Boston: Butterworth-Heinemann. Hsieh, C. (2005). Strengths and weaknesses of qualitative case study research [online]. Available at: http://140.109.171.170/download.php?filename=51_544bcf1c.pdf&dir=paper&title=%E5%85%A8%E6%96%87%E4%B8%8B%E8%BC%89 [Accessed 29 May 2013]. Isaac, S. and Michael, W. (1997). Handbook in research and evaluation: A collection of principles, methods, and strategies useful in the planning, design, and evaluation of studies in education and the behavioral sciences. (3rd Ed.). San Diego: Educational and Industrial Testing Services. Jensen, J. and Rogers, R. (2001). Cumulating the intellectual gold of case study research. Public Administration Review, 61(2), pp. 235-246. Kalton, G. (1983). Introduction to survey sampling. Newbury Park, CA: Sage University Publications Kraemer, K. (1991). Introduction. Paper presented at The Information Systems Research Challenge: Survey Research Methods. Lazaraton, A. (1995). Qualitative research in applied linguistics: A progress report. TESOL Quarterly, 29(3), pp. 455-472. Liu, Z., Fu, C., Wang, W. and Xu, B. (2010). Research Prevalence of chronic complications of type 2 diabetes mellitus in outpatients-a cross-sectional hospital based survey in urban China. Health Qual Life Outcomes, 8, p. 62. Marrelli, A., (2004). Surveys. Performance improvement, 43(10), pp. 38-43. Mathiyazhagan, T. and Nandan, D. (2010). Survey research method. Media Mimansa, pp. 34-40. Merriam, S. (1988). Case study research in education: A qualitative approach. San Francisco: Jossey-Bass. Murphy, J. and Addis, L. (2008). Cast your vote: the results are in. Internet: Rev Optom [online]. Available at http://www.revoptom.com/index.asp?page=2_14071.htm [Accessed 29 May 2013]. Sanglimsuwan, K. (2011). Review and comparison of sampling designs and survey methodologies in environmental health research. Bangkok University [online]. Available at: http://www.bu.ac.th/knowledgecenter/executive_journal/july_sep_11/pdf/aw5.pdf [Accessed 30 May 2013]. Silverman, D. (2005). Doing qualitative research. 2nd ed. Sage Publications: London. Silverman D. (2006). Interpreting qualitative data: methods for analyzing talk, text and interaction. 3rd ed. Sage Publications: London. Tourangeau, R. (1999). Interdisciplinary survey methods research. In M. G. Sirken, D. J. Herrmann, S. Schechter, N. Schwarz, J. M. Tanur, & R. Tourangeau (Eds.), Cognition and survey research. New York: John Wiley and Sons. Wild, J. and Ehlers, A. (2010). Self-study assisted cognitive therapy for PTSD: A case study. Eur J Psychotraumatol. Yin, R. (2009). Case study research: design and methods. 4th ed. Sage Publications: London. Read More