GGR271 Final Study Guide.doc

31 Pages
Unlock Document

Matti Siemiatycki

15/04/2013 17:17:00 GGR271H1 Introduction to Data sources and Methods – Week 2 Epistemological positions- a branch of philosophy concerned with what constitutes knowledge and how knowledge is acquired. • Positivism o Science is the basis of all the laws that explain human behaviour o Only scientific testing can lead to knowledge o Theory generates hypotheses that can be tested o Science is value free o Distinction between science and normative statements o Advocates using the methods of the natural sciences in the study of social reality • Intrepretivism o Seek to understand human behaviour o Cannot separate the researcher from the researched  Understand human role in their own surroundings o Research should be action oriented: research informs and shapes social change o Requires the social scientist to grasp the subjective meanings that people attach to their actions and behaviours Ontological Positions –a branch of philosophy concerned with the nature of reality: whether social entities are considered objective and external to social actors, or are they social constructions built up through perceptions and actions of these actors • Objectivism o Social phenomena exist beyond the reach of individuals o Organizations have rules and regulations that possess an external reality • Constructivism o Social phenomena and their meaning produced through interactions o No pre-existing order o Researcher presents only reality of the world Nomothetic explanations • Used by those who subscribe to the natural science/ quantitative methods • An explanation that applies to all of humanity • Has to satisfy 3 criteria o Correlation  The proposed cause and the proposed effect vary together o Time order  Their proposed cause must precede the effect in time o Non-Spuriousness  Try to determine whether some third factor is the cause of the correlation Qualitative researchers aim for less generalizing definitions • Seek to produce a rich description of a person or a group • Rather than to discover general laws and principles o These are called idiographic explanations  Detailed story of the people studied that is based on empathetic understanding Data Sources • Interviews/questionnaires, letters, corporate documents, etc. Structured Observation- technique where the researcher follows explicitly formulated rules for the observation, categorization, and recording of behaviour • Strengths o Examine underlying issues not found with surveys o Examine the gap between stated and actual behaviour • Weaknesses o Ethical considerations o Inaccurate coding imposed on human behaviour through error  Perhaps and inappropriate observation schedule was selected? o Not able to understand intentions behind observed behaviour o Generates many small bits of data that can cloud an analysis of themes Primary Data – data collected directly by researcher • Advantages o Can tailor data collection specifically to research question o Ensure quality required in data collection • Disadvantages o Time consuming o Can be costly o Run into ethical considerations, its obtrusive into peoples lives Secondary Analysis of Data – the analysis of data collected and produced by other researchers. • Advantages o Enable research in areas that you can’t be accessed through surveys etc. (i.e. ethics) o Reanalysis often offers new insight into the original study o Unobtrusive, lower chance of harming participants of a study o Less time consuming, and less costly o Data is usually high quality o Can help with cross cultural analysis • Disadvantages o Cannot control reliability and validity o Can be vast and complex o Must be care not to fall into ecological fallacy  The idea that you can characterize a geographic area as impoverished and dangerous because of the high number of visible minorities (for example).  Rather than interviewing individuals, it’s the process of looking at commonalities between the demographics in the geographic areas, not the commonalities between survey results. Evaluate the quality of evidence • Authenticity: is the evidence genuine and of unquestionable origin? • Credibility: Is the evidence free from error or distortion? • Representativeness: is evidence typical of its kind, and if not is the extent of its uniqueness known? (I.e. extreme cases) • Meaning: is the evidence clear and comprehensible? Hierarchy of the quality of sources 1. Peer reviewed data 2. Non peer reviewed, but academic data 3. Expert reports from governments, NGO’s etc. 4. Factual articles from magazines, newspapers, etc. e.g. The New York Times 5. Opinion pieces such as editorials, and blogs. Basic research: • Research to develop theory and to gain knowledge in general, academic and social interest Applied research • Done (perhaps on a salary – so you have to worry about bias) to answer a specific question for a person, institute, school, etc. The Research Design Process - week 3 The Research Process 1. Research question 2. Review literatures 3. Devise design 4. Select site and research method 5. Review ethics and apply for approval 6. Administer research and analysis 7. Write up findings Selecting research questions • Broad enough to be interesting, narrow enough to be doable • Make sure data is accessible • Qualitative research will usually be inductive o Based on observations in society Research design • A framework for collection of data and its analysis Research Designs • Cross sectional surveys o Lots of data on more than one case collected at A SINGLE POINT IN TIME o This body of data lets comparison take place between the cases o Interested in the variation between comparison o Like a snapshot taken of a group at one point in time o Usually a collection of quantitative data o Data collected on 2 or more variables to be able to see association between the 2 o Due to the time restraints it is hard to see which variable is influencing which, just that there is a correlation o Uses questionnaires  Structured through many mediums – telephone, mail, in person  Often analyzed statistically • Longitudinal Design o Cases examined at a particular, then at least once again at a later time o Is a much better design to infer a causality between 2 variables o More costly and more time consuming  2 types of longitudinal designs  Panel • Sample group is chosen and studied on at least two occasions • Always the same participants • Has strict rules for new entries into the program • Prolonged participation can cause ‘panel conditioning effect’ o Where being in the study has an impact on how the participants act  Cohort • A cohort is selected who have similar characteristics (like a graduating class) and sampled on at least 2 occasions • Not necessarily the same participants each time, but are always randomly sampled from within the cohort.  Both use the repeated survey structure usually  Both deal with the problem of sample attrition, losing members through death, moving away, withdrawal • Case Study design o Detailed intensive analysis of one single case o Usually qualitative data collection  Thus inductive analysis is used which provides idiographic explanations and themes The critical case  Clearly stated hypothesis  The study will give a better understanding of on which grounds the hypothesis stands, or falls apart The Unique case  Common in clinical cases  Somewhat extreme uncommon behaviour  Sometimes offers a chance to challenge what are considered accepted views. The revelatory case  A study of newly available material that had been hidden/lost/ locked away before The Convenience case  Nothing special happening  Yet, still offers valuable capacity to study important links between social actors, and inform theory o Case Study issues  Research developed cannot be generalized, can’t be applied to other aspects of society – it is specific to the case only  Case study researchers argue that a valid picture of one case is more valuable than a less valid picture of many cases  Used to generate and inform theory Research can be evaluated based on 3 criterion that measure the variables in the study 1. Reliability: would the same results be achieved if the same research technique were administered several times to the same subject? 2. Replicability: if someone else carried out the same research would they get the same results? (ideally yes!) 3. Validity a. Measurement validity i. Whether the social indicators within the research actually measure what they say they do b. Internal Validity i. The causality hypothesized. How confident are you in it? Does in the independent variable really affect the dependent variable? c. External Validity i. Are the findings applicable to the world outside of the research setting? Researchers who are more focused on qualitative data, have suggested that their studies not be evaluated with “validity” but rather with a proposed measure named “trustworthiness” which would judge the following: o Credibility – no personal bias o Do the findings apply to some other people? o Is the work dependable and consistent over time? o Would other researchers reach the same conclusions?  These have many parallels with the measurement of validity Once a design is selected the next step is to collect data • Through either structured surveys and interviews, or unstructured surveys and participants Sampling Theory & Practice - Week 4 Why survey? • Facts, behaviours, preferences, predictions • Supplements data • Everyone surveys! Main concepts of research • Reliability & Precision: will it have the same result consistently? • Validity & Accuracy: Does the test measure what it claims to o You need BOTH for a quality research project Sampling Probability Sampling Simple Random Sample • Most basic probability sampling • Each unit of the population has an equal probability of inclusion in the sample • No opportunity for bias, and does not depend on the individuals availability Systematic Sample • Using random number pick a sampling interval • Make sure there is no inherent ordering within the sampling frame Stratified Random Sample • Divide population based on criterion • Take sample (random or systematic) from each of the divided levels • Like dividing all the universities in Ontario into groups, and then randomly sampling from each group • Select samples in proportion to the population • Feasible only when easy to differentiate where to divide levels • Still technically representative Multi stage cluster sampling • E.g. Divide country into regions o Randomly select (sample) a number or regions  Randomly sample number of subjects within randomly selection regions • Still representative if done properly (random) Probability Sampling (all of the above) • Members of a population are included in a survey by chance • Intended to be representative • Can use phone book, postal codes, table of random numbers • Good! Because the findings are externally valid. They are representative of the whole population and thus are generalizable to the rest of the population • Sources of Error o Sample size  Bigger sample can increase precision most often, but cannot guarantee anything o Non Response  ‘Response rate’: percentage of sample that actually participates in the survey  Non-response can be caused by unavailability, refusal to participate, unfinished questionnaires  Response rate often as low as 59%  Those who don’t response still have to be examined as much as possible – try to find the reason for refusal, etc.  Can adjust sample size beforehand to take into account the threat of non response o Heterogeneity of Sample  Heterogeneous (varied) populations should have a large sample because of the likelihood of very different answers.  Homogenous samples (like high schools) can have smaller samples because likely all answers will be similar Purposive Sampling • Purposely choose respondents who are thought to be relevant to the research topic • Not often representative Snowball Sampling • Used when sampling hard to reach/ hidden populations o I.e. gay lawyers in New Brunswick (random sample would be impossible • Not representative- cant apply generally • Researcher contacts small amount of gay lawyers and pulls in other members of the population through the initial subjects Theoretical Sampling • Combines convenience and snowball sampling • Researcher simultaneously collects and analyzes data • Then decides what data to collect next and where to find it • Data collection is influenced continuously by emerging theory • Data collection continues and evolves until you reach “theoretical saturation” which means o No new data is emerging o Topic is extremely well developed in all areas o The relation ship among categories is well developed and well argued Convenience/opportunity Sampling • Uses elements readily available to researchers o I.e. students or family • A good model to dry run a newly created system of question/scales/surveys Quota Sampling • Like stratified random sampling • Instead of random sample from each strata, researcher dictates a quota for each strata o Ask the first people the bump into to fulfill their quota • Not random – but has been argued to be quite representative Questionnaire/Interview based research - Week 5 Social surveys • A set of techniques that are used to gather data by direct communication with individuals • Can range from loose to very tight set of questions • Either quantitative or qualitative • Many ways to deliver it Structured Interview • Administration of a standard interview • Exact same set of question delivered the same way each time • Promotes standardization in asking of questions and recording of answers • Reducing error from variation in delivery of questions • Highlights actual variation between respondents • Survey/Interview error o Poorly worded questions o Bias or tone when questions are asked o Misunderstanding, forgetfulness, recording of response wrong Construction a questionnaire • Close ended questions (multiple choice or scales (likert) from 1-5) o Reduces variability by both interviewer and interviewee o Faster to process • Open ended questions o Hard for interview to record everything said o Room for misinterpretation o More error • Likert Scale o List a statement and have the respondent indicate their level of agreement o No more than 7 different option • Question ordereing o Start general and move towards specifics or vice versa o Personal questions at the end generally o Mix positive and negative questions of a similar topic together o Remember that one question may prompt how the respondent answers following questions o Don’t Use: double negatives, very technical language, and hypothetical questions • Coding o Assign labesl to recurring themes or categories  i.e. hostile to outsiders o codes are given so they can be stored quantitatively Interviewing In Practice - Week 6 Interviews • More open ended than surveys • Statistics are replaced with benchmark figures • Interview theoretically useful respondents • Interviews have characteristics that enable comparison Unstructured Interview • Interviewer has AT MOST a memory aid- more of a conversation Semi-structured Interview • Interviewer has a list of questions/topics but has wiggle room • Respondent can go off on tangents if wanted When Interviewing… • Don’t judge: make people feel comfortable answering • Let people talk – even if they ramble • Select a comfortable semi-private venue • Tailor the interview o Use descriptive questions (prompts) to lead them down the right path if needed • Give them info about the research project before the interview • Have an interview guide (structured or not) • Let your topics be fairly broad • Ask follow up questions to responses for more detail and information • If you cross check for verity (this is okay) just make sure y
More Less

Related notes for GGR271H1

Log In


Join OneClass

Access over 10 million pages of study
documents for 1.3 million courses.

Sign up

Join to view


By registering, I agree to the Terms and Privacy Policies
Already have an account?
Just a few more details

So we can recommend you notes for your school.

Reset Password

Please enter below the email address you registered with and we will send you a link to reset your password.

Add your courses

Get notes from the top students in your class.