COMM 301: Empirical Research in CommunicationSurvey / interview method FunctionsSurvey / interview method AdvantagesSurvey / interview method DisadvantagesSurvey / interview method Disadvantages (cont.)Survey/Interview DesignsSurvey/Interview Designs (cont.)Functions of the questionnaireQuestion typesGuidelines for developing questionsSlide 11Slide 12Slide 13Slide 14Slide 15Guidelines for developing questions (cont.)11COMM 301:Empirical Research in Communication Lecture 8Kwan M Lee22Survey / interview methodFunctions•Both survey and interview are self report techniques.•Able to serve two functions–Description •document current conditions•E.g., Opinion poll–Analysis•describe and explain why conditions occur•What are underlying causes and relationships?33Survey / interview methodAdvantages–high ecological isomorphism•Data gathered in real world situations–reasonable cost of data collection–access to broadly distributed population•Can cover broad location–access to subjective information•respondents’ private thoughts44Survey / interview methodDisadvantages–no control over independent variables•hard to rule out alternatives  IVs are not under control; they are simply measured–lack of temporal ordering•correlational research -- focus on relationships between variables•not causal relationships  because it cannot control time order–instrumentation bias•wording and organization of question items can affect the data collected – e.g., order effect•onus to develop questions well55Survey / interview methodDisadvantages (cont.)–requires respondents’ willingness•respondents may provide inaccurate or incomplete data•can be checked by survey instruments – e.g., parallel questions to check consistency–requires respondent recall–prestige bias•Respondents selected because of their specific SES may provide dishonest or biased answers–declining response rate•People hate it.66Survey/Interview Designs•Types (in terms of time span)–Cross-sectional vs. longitudinal–Cross-sectional design•data collected from a single point in time•create a description of the conditions at that time•equivalent to a still photograph, snapshot77Survey/Interview Designs (cont.)–longitudinal: data collected from a sample across a period of time•trend study–data collected from different samples selected from a single population at different time points–evaluate changes in the population•panel study–data collected from same sample at different time points–follow the sample over time to evaluate changes–examples: genetic predisposition to violence, television viewing and violence–vulnerable to which threat to validity?»attrition88Functions of the questionnaire•Designing survey instrument, aka the questionnaire–need to accomplish two tasks•convey information to respondent clearly and concisely, so that they understand what is being asked•provide a way for respondent to answer accurately and concisely99Question types•Question types, and their relative tradeoffs–closed ended•response options given by the question, respondent chooses the appropriate response•easier data input and analysis•lack detail–open ended•no predetermined response options, respondent can elaborate details•can collect data previously unknown to researcher•data analysis more problematic1010Guidelines for developing questions•1. Make questions clear–Meaning: no ambiguity•example of ambiguous, broad question:–“What do you think of our promotion?”•need to specify as much as appropriate–“what do think of the speed of rebate? ~ easiness of rebate instruction? ~ ~ ~–simple, understandable language•avoid complex multi syllabi words, specialized terms (without clarification), and negative wording•Ninth-grade level–Avoid using words with multiple meanings•e.g. “hot”: Is your instructor hot?1111Guidelines for developing questions•2. Make questions concise–Good questions are short, yet clear–may consider breaking long questions into multiple questions1212Guidelines for developing questions•3. Include complete instructions–provide instructions about how to select items•rank order, choose only one, choose every items applicable?–clear instructions for filter questions•e.g. asking about email use for people without email–filter question: “Do you have access to email”  Yes  No (skip to Question 26)1313Guidelines for developing questions•4. Be realistic–About respondents’ abilities to recall•e.g. how many hours did you watch TV during the last year? ? ?•5. Focus on purpose of survey–Control the length of a survey–Do not include items if they are not relevant to your research1414Guidelines for developing questions•6. Avoid double barreled questions–a single question that asks for more than one response•e.g. “Do you rely on the Internet to communicate with people and purchase merchandise?”–Use of conjunctions (and, or, but…) are yellow flags indicating possible double-barreled questions1515Guidelines for developing questions•7, 8. Avoid biased wording and leading questions–biased wording that can shape a response•e.g. “Do you read newspapers, or just watch television?”–other bias-prone words (words about ideology, political party…)•E.g., Do you agree with the Republican’s tax cut proposal?–leading questions that suggest specific responses•e.g. “Like most people, do you support anti-abortion?•e.g. “Don’t you support the anti-abortion proposals?”–double bind question•implies affirmative response regardless of how response•e.g. “Do you still cheat?”1616Guidelines for developing questions(cont.)•9. Avoid threatening questions–questions that cause unease or embarrassment•sensitive topics–sexual behavior, gambling, substance abuse, etc.•personal and private information–age, income,…–possible solutions•ask for equivalent answer if possible–e.g. year of birth instead of age•use categories, with an artificially low initial category–e.g. income •use random techniques–Number of cheating - A: 4; B:1; C:2;