Study Guides (247,982)
Canada (121,201)
SOAN 2120 (93)
D Walters (37)


21 Pages
Unlock Document

Sociology and Anthropology
SOAN 2120
D Walters

SOAN2120 FINAL EXAM REVIEW LECTURE NOTES PURPOSE OF THIS COURSE (1 Question)  The purpose of this course is to provide an introduction to research methodology in the social sciences  If you have one of each use a CROSS TABULATION TEXBOOK: SAMPLING (9 Questions) Types of Sampling Quantitative methods primary goal is to get a representative sample small collection of units from a larger collection of units sampling based on theories of probability Non-probability sampling  Rarely determine the sample size in advance and have limited knowledge about the larger population from which the sample is taken selects cases gradually, with the specific content of a case determining Types of Sample Principle Haphazard Get any cases in any manner that‟s convenient Quota Get a present number of cases in each of the several predetermined categories that will reflect the diversity of the population using haphazard methods Purposive Get all possible cases that fit particular criteria, using various methods Snowball Get cases using referrals from one or a few cases, and then referrals from those cases, and so forth Deviant Case Get cases that subsequently differ from the dominant pattern (A special type of purposive sample) Sequential Get cases until there in no additional information of new characteristics (often using other sampling methods)  Haphazard can produce ineffective, highly in representative samples and is not recommended; sample that seriously misrepresents the population cheap and quick, but worse than no sample at all example: street interviews  Quota identifies a relevant categories of people and then decides how many to get in each category; difficult to represent population accurately ensures that some differences are in the sample  Purposive is used in situations in which an expert uses judgements in selecting cases with a specific purpose in mind may never know if cases represent actual population often used in exploratory of field research appropriate in 3 situations: to select unique cases that are especially informative, to select members of a difficult to research population, wants to identify certain types of cases for an in-depth investigation  Snowball is a method for identifying and selecting the cases in a network multistage technique: begins with a few cases and spreads out on the basis of initial cases crucial feature is that one is unit is connected (indirectly or directly) to another represents such networks by drawing a sociogram  can be used in combination with purposeful sampling  Deviant is used when researcher seeks cases that differ from the dominant pattern or characteristics of other cases goal is to locate a collection of unusual, different, or peculiar cases that are not representative of the whole  Sequential sampling similar to purposeful sampling with one difference, as they continue to gather new cases until new information/diversity is full Probability Sampling Sampling element: unit of analysis or case in a population Population: large pool (sometimes called universe) and can be anything that is measured Target population: specific pool of cases that he or she wants to study Sampling ratio: the ratio of the size of the target population Sampling frame: developing a specific list that closely approximates all the elements in the population almost always inaccurate and therefor bias towards the population Population parameter: any characteristic of a population, determined when the elements of the population is measures Statistic: information from the sample used to estimate population parameter Sampling error: random sampling lets a researcher calculate the relationship between the sample and the population (size of sampling error) the deviation between sample results and the population parameter due to random process Simple random sample: is both the easiest random sample to understand and the one on which others are modelled researcher develops and accurate sampling frame, selects elements from the sampling frame according to a mathematical random procedure, then locates the exact element that was selected for inclusion in the sample (random number generators) Sampling distribution: a distribution of different samples that shows the frequency or different samples from many separate random samples. Central Limit Theorem: tells us that the number of different random samples in a sampling distribution increases toward infinity, the pattern of samples and the population perimeter become more predictable with a huge number of random samples, the sampling distribution forms a normal curve and the midpoint of the curve approaches the population parameter as the number of samples increase  Random sampling does not guarantee that every random sample perfectly represents the population; it just means that most random samples will be CLOSE to the population most of the time, and that one can calculate the probability of a particular sample being inaccurate. Estimate the chance the sampling is off (size of sampling error) by using information from the sample to estimate the sampling distribution, and combines this information with knowledge of the CLT to construct confidence intervals Confidence intervals: a range around a specific point to estimate the population parameter. Range is used because researcher cannot predict exact point, but rather with a level of confidence (eg. 95%) that the true population lies within a certain range calculations of sampling errors are based on the idea that the sampling distribution that lets a researcher calculate the sampling error and confidence interval (using CLT) Calculating samples Step 1: Number each case in the sampling frame in sequence (ex. alphabetical order/1-40) Step 2: Decide on sample size Step 3: For simple random sample locate a random number table. Count the largest number of digits needed for the sample (ex, 50 cases = 2 digits) …….. the fuck this is too complicated I quit. LECTURE NOTES SAMPLING (3 Questions) Nonprobability sampling (QUALITATIVE) no math!  Doesn‟t know sample in advance Probability Sampling (QUANTITATIVE) uses mathematical theory  has some knowledge about the population but they are abstract and variable  Goal is making inferences about an unknown populations  Generally have boundaries (ex. geographic location) Simple random sample (SRS): assign a number to each element, use a random number generator to sample the people in the same way, resulting in equal probability. benefits: get the sampling distribution through the CLT and you can make inferences about the population, all statistical inference is invalid without random sampling Systematic sampling: decide on sample size, get sampling interval by dividing population by sample size (ex. population (1000) ÷ sample size(100) = sampling interval (10)) start with your sampling frame (after numbering each element) and pick one random number after every 10 thereafter difficult to context to group Stratified sampling: divide the populations into sub-populations (stratum) and the sample within the stratum (systematically or randomly). Stratums are chose by the researcher dependent on the research question. There can be multiple stratums want to guarantee a certain number of the people from each stratum are adequately represented in the sample so you must apply weights for probability guarantees stratums are represented in the survey Cluster sampling: identify clusters (ex. city blocks, households ect.) then SRS clusters and then SRS units within the clusters (ec. Cluster-city blocks, units- household, units within units-individuals) multiple samples leave room for many systematic errors Each small rectangle is a Each circle represents a region city block Randomly pick some city Inside regions are city blocks blocks Take a SRS of the Randomly sample individuals within these individuals within each city block blocks Stratified Cluster Comparing the Methods Sampling decisions are usually a trade-off between accuracy and efficiency Stratified is more representative Cluster is more cost effective of the entire population VERSUS TEXTBOOK SURVEY RESEARCH (10 Questions) Question wording principles: keep it clear, keep it simple, and keep the respondents perspectives in mind. Good survey questions give the researcher valid and reliable measures, as well as help respondents understand the question and their answer is meaningful.  Avoid jargon, slang and abbreviations  Avoid ambiguity, confusion and vagueness causes inconsistencies in how different respondents assign meaning to and answer the question source of ambiguity in the use of indefinite words or responses (eg. what does regularly mean?)  Avoid emotional language words with strong emotional connotations can change how respondents answer the question use neutral language  Avoid prestige bias titles or positions within society carry prestige or status, issue slinked to people with high status can affect the way people respond (ex. bias)  Avoid double barreled questions make each question about ONE topic double barreled consists of two questions joined together which makes respondents answers ambiguous, thus researcher can‟t be sure of meaning  Do not confuse beliefs with reality do not confuse respondents beliefs with what researcher is measuring peoples beliefs about a relationship among variables are distinct from an actual empirical relationship  Avoid leading questions make respondents feel like all responses are legitimate and do not let them become aware of what the researcher wants  a “leading/loaded” question is the one that leads the respondent to choose one response over another by its wording, and can be used to get either negative or positive reactions  Avoid asking questions beyond respondents capabilities asking something few people know about frustrate respondents and produces poor-quality responses phrase questions in terms of the way respondents think  Avoid false premises do not begin a question with a premise with which a respondent may not agree, then ask about choices regarding it better to explicitly ask respondent if premise is true, then ask for preference  Avoid asking about intentions in distant future avoid asking what people might do under hypothetical situations far in the future, responses are poor predations of behavior  Avoid double negatives they are confusing. Just don‟t. arise when respondents agree and disagree with a statement  Avoid overlapping/unbalance response categories make the response categories or choices mutually exclusive, exhaustive and balanced mutually exclusive: means response categories do not overlap exhaustive: means every respondent has a choice balanced: offering bipolar opposites at each end of a continuum Types of Questions  Threatening questions are part of a larger issue of self-presentation and ego, surveys may ask sensitive questions and respondents may be reluctant to answer truthfully. May be ashamed or embarrassed to answer truthfully. may underreport or self-censor reports of behavior or attitudes they wish to hide or believe deviate from social norms (disability, deviant behavior, financial status) should ask threatening questions only after a warm up, when an interviewer has developed rapport and trust with the respondents should tell respondents they want honest answers can phrase questions in an “enhanced” way to provide a context that makes it easier for respondents to give honest answers embedding a threatening response within more serious activities, it made me made as less deviant  Socially desirable questions occur when respondents distort answers to make their reports conform to social norms.  people tend to over report being cultured, giving money to charity, having a good marriage, loving their children, ect. reduce social desirability by phrasing questions in a way that make norm violation appear less objectionable and that presents a wide range of behavior as acceptable can also offer multiple response categories that give respondents “face saving” alternatives  Knowledge questions can be threatening because respondents don‟t want to appear ignorant can measure opinions better if they first ask about factual information, because many people have inaccurate factual knowledge some simple knowledge questions aren‟t answered accurately (how many people live in your home) can be worded so respondents feel comfortable saying they don‟t know anything about it  Contingency questions are the two or more part question in which the answer to the first part determines which of the two different questions the respondent next receives. It selects respondents for which the next question is irrelevant, sometimes called “skip” or “screen” questions on the basis of the first question, the respondent is instructed to go to another or skip the next question  Open VS Closed questions: an open ended (unstructured, free response) question asks a question to which respondents can give any answer. A closed ended (structured, fixed response) question both asks a question and gives the respondent fixed responses from which to choose choice to use either type depends on the purpose and practical limitations of a research project large scale surveys have closed ended questions because they are quicker and easier for respondents and researchers sensitive topics may be more accurately measured with closed question Advantage OPEN ENDED CLOSED-ENDED  Unlimited # of answers  Quick and easy  Detailed answers and can  Easier to compare different clarify responses answers  Unanticipated findings can  Answers are easier to code be discovered and analyze  Permit adequate answers to  Respondents choices can complex issues clarify question meaning for  Permit creativity, self- responses expression and richness of  More likely to answer about detail sensitive topics  Reveal respondents logic,  Fewer irrelevant or confused thinking process and frame questions of reference  Less articulate/illiterate are not at a disadvantage  Replication is easier Disadvantage  Different degrees of detail in  Suggest ideas respondents answers would have not otherwise had  Responses may be irrelevant or buried in  Respondents with no useless detail opinion or knowledge can  Comparisons and statistical answer anyway analysis may be very difficult  Respondents can be  Coding responses is difficult frustrated because their answer is not a choice  Articulate and high literate respondents have  It is confusing if too many advantage answers are offered  Questions may to be general  Misinterpretation of a for those who lose direction question can go unnoticed  Responses are written,  Distinctions between difficult for interviews respondents answers may  Great amount of respondent be blurred thought, time and effort are  Clerical mistakes or making necessary the wrong response is  Respondents can be possible  Force respondents to give intimidated by questions  Answers take up a lot of simplistic answers on space in questionnaires complicated issues  Force people to make choices they would not make in the real world  Non-attitudes and the middle positions: surveys debate whether to include choices for neutral, middle and non-attitudes. Two types of errors can be made: accepting the middle choice or no attitude responses when respondents hold a non-neutral opinion, or forcing respondents to choice a position on which they have no opinion about it is usually best to offer a non-attitude choice because people will express opinions on fictitious issues, objects or events. By offering this choice researches identify who hold middle positions and who doesn‟t give a fuck. many will answer a question if “no opinion” choice is missing, but will chose a “don‟t know” when it is offered, or say they do not have an opinion if asked these respondents are called floaters because they “float” from giving a response to not knowing. Researchers screen them out using quasi or full filter questions standard format: does not offer a “don‟t know” choice, a respondent must volunteer it quasi filter: offers a respondent a “don‟t know” alternative full filter: is a special type of contingency question, it first asks if respondents have an opinion, then ask those who do to state it. Agree/Disagree, Ranking or Ratings it is best to offer respondents explicit alternatives rather than these types. Researchers create bias if wording gives reason for choosing one alternative Interviewing  Having interviewers periodically using probes to ask about a respondents way of thinking is a way to check whether respondents are understanding the questions as researcher intended could shape respondents answer or force answers when respondent does not have an opinion or information  Flexible or conversational interview in which interviewers use many probes can improve accuracy on questions about complex issues on which respondents don‟t clearly understand basic terms about, or which they are having difficulties expressing their thoughts  Partially open questions: a set of fixed choices with a final open choice of “other”, which allows respondents to offer an answer that the researcher did not include. Stages of an Interview  An interview proceeds through stages beginning with and introduction and entry shows authorization and reassures and ensures cooperation from the respondent can explain why respondent was chosen and not a substitute  .main part of the interview consists of asking questions and recording answers, using the exact working on the questionnaire, no words omitted or changes of phrasing. Asks these questions in order, without returning or skipping any unless it specifies it.  .goes at a comfortable pace and gives nondirective feedback to maintain interest  For recording answers, interviewer must listen carefully, have legible writing and must record what is said in verbatim without correcting grammar or slang, never summarizing or paraphrasing  Interviewer knows how and when to use probes, which is a neutral request to clarify an ambiguous answer, to complete and incomplete answer or to obtain a relevant response.  Types of probes: three to five second pause, non-verbal communication (tilting fo the head, raised eyebrows, eye contact), repeat question or reply then pause, ask neutral questions “any other reasons?” “could you explain more?”  Last stage is exit, where interviewer thanks respondent and leaves.  Then goes to a quiet, private place to edit questionnaire and record details such as date, time and place, thumbnail sketch of respondent and interview situation, respondent‟s attitude, any unusual circumstances. Notes anything disruptive that happened during the interview. Records personal feelings and anything that was suspected. Types of Surveys Features Mail Web Survey Telephone Face-to-face Questionnaire Interview Interview Cost Cheap Cheapest Moderate Expensive Speed Slowest Fastest Fast Slow to Moderate Length Moderate Moderate Short Longest Response Rate Lowest Moderate Moderate Highest Probes possible No No Yes Yes Specific No Yes Yes Yes respondent Question No Yes Yes Yes sequence Only one No No Yes Yes respondent Visual No No No Yes observation Visual Aids Limited Yes None Yes Open-ended Limited Limited Limited Yes questions Contingency Limited
More Less

Related notes for SOAN 2120

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.