Class Notes (811,155)
Canada (494,530)
COMM 2002 (65)

COMM2001 October 7, 2013.docx

10 Pages
Unlock Document

Carleton University
Communication Studies
COMM 2002
Heather Pyman

Questionnaire Construction •When we look at studies that are done, we should be able to bring critical perspectives •Make linkage between level of education and daycare stuff • Questions about different choices of daycare •What they are actually doing •Look at differences between males and females •What kind of daycare choices that are made •Look at the effect of X on Y •Mitigate effects and variables in different scenarios •Thinking about reliability and responsibility •If we were to ask questions in another scenario, would we get the same kind of pattern and results? •Is there a pattern and can we say with some sort of confidence that ___ is an indicator of ____? •Is it reliable? When/where was it collected? •Does time have an impact between things? Time of the year? Are there elements with effects on results that we don't actually see and should consider when assessing the value of the research that we do? •Get to accuracy of how you calculate statistics •Point out how data is collected and where the samples are coming from •Think about the relationship between education and daycare choices -> samples •Consider the following: o Date o Size o Mode (how it was done -> online poll) •What may be issues that may be related to findings that you have and the mode? (ex. If it is an online polls, what factors?) o -> access to computer o The audience •Research Questions •Need to know our audience, who we are targeting, is it something that we need to know? •Canadians with higher level of education -> are we talking about Canadian, provincial population •What is our research question and where are we getting our answers from? When designing questionnaires: •Remember your research questions •Decide exactly what you want to find out o Keep in mind analytical areas as well to understand research questions o Analytical variables •Imagine yourself as a respondent o One of the problems we get in survey construction is that the researchers do not think about who is answering this question • Language, tone, complexity, specific kinds of questions asked in survey, have to take into account who the person is answering it o Is it general audience? Youth? Whole population? o Sensitivity of the education o Make sure that the person who is answering the questions can actually answer them o The person may answer that question for you Mode Choices • Different modes o Telephone vs. person/mail vs. internet • Advantages/disadvantages • Emphasize that each of them is a potential mode, they all have advantages and disadvantages, we have to choose the one that fits the best for client and the amount of money that we have • If we were to compare the telephone versus in person interviews, you can think of some of the advantage and disadvantages of doing it • Telephone o Advantage: Distance, reach, coverage is limitless, Convenience, Quicker, takes less time, easy to supervise, inexpensive o Disadvantage: Limited landlines • In person interview o Advantage: make connection with respondent, visual cues and ability to explain when people don't understand the question, intercept surveys, Visual cues, higher completion, location specific o Disadvantage: Personal bias • Self completion: Compare mail surveys versus internet surveys • Internet: o Coding is done automatically o Access to large groups o Cost effective o Issue of confidentiality • Mail: o Cost time o Time can affect reliability • Respondents are told in an academic situation that the researcher needs to connect your responses to you photographically • When they say data is anonymous, they remove your specific name or address from the research Constructing Questionnaires • Identify yourself and introduce the research o Who you are/who is funding the research/the purpose/respondent selection/contract # • Provide clear instructions o How long the interview is o Address issues of confidentiality or anonymity o Try to disguise some of the research in market research o Try as many different techniques as you can to not do things • Consider question order o Start with easy questions first and finish with demographic questions o Once they start survey, they get to a draw o Make sure that the person who is answering the question has ability or criteria to answer the questions • Use filter questions where appropriate o When asking about daycare, or no experience with daycare, it will greatly influence the questions that you have when people are looking for different sources of daycare, o Need to have filter questions (Have you currently or have you been looking at day care, if no, skip to ten) • Ensure quality interviewing o Make sure everything follows into the same patter so everyone has the same experience o Make sure that the questions are not interpreted for you o Slight changes in wording can change meaning o Need to determine whether you are going to allow if the interviewer can go further o Is this a possibility o General elements you need to consider o Determine ordering if question and making sure that all the instructions are written down or given to interviewer so everyone has same experience to the actual questionnaire • Need to be sure that you are giving respondents clear instructions • Make sure that everything that you think of has a possible question area that you covered • Are you asking them to circle answer? Tick the box? Do they need to have prerequisite questions? • Need to fully explain to respondent how to answer those particular questions Question types • Closed vs Open ended Questions o Advantages/disadvantages o Why would you or not support universal daycare? (Open ended question) -> They can express themselves in how they are interpreting the question rather than trying to fit how they are feeling in a close set o As opposed to I support universal daycare because it makes it more affordable o Disadvantage is that you have to read every response o Very time consuming o No spontaneity o Are answers the same elements of something? o Are these all part of a new code to be created? o Advantage is that as long as you have a mutually exhaustive list, it is quick, easy and gives you consistent results • Usually most surveys are closed because they are difficult to code, most of the time, some open ended questions • These are concepts to be measured, how would you measure them? What are we thinking about the issue? • Types of questions • Attitudinal - Beliefs/Values/Knowledge • Can you measure them like this: o Behavioural - Actions o Demographic - Features o Come up with a mutually exclusive, exhaustive list • Usually have a code in your coding scheme, all the possible answers and you can create a code where someone can write in the box that they do not want to write in the box • Are these going to be analytically helpful? Writing questions: Specific Rules: • Avoid o Ambiguous terms o Technical, jargon and acronyms o Long questions o Double-barrelled questions • Needs to be specific • Make sure you ask only one question o Very general questions • Need a frame of reference o Leading questions o Negative terms ('not', 'never') • Do not use double negatives, especially in interviews where someone cannot look at text • How easy are elements to creep into the way we are writing questions • When we are writing, we know what concept we are trying to measure • Make sure that the respondents have the requisite knowledge • Symmetry between closed questions and answers • Create a balance between positive and negative responses to a question • Don't rely on respondent's memory • Include a "Don't know" option • Aware off validity, how likely the can remember (ex. Who did you vote for) Common mistakes when designing surveys • Excessive use of open questions • Excessive use of yes/no questions • No instructions about how to indicate answers (tick box, circle, delete?) • Overlapping response categories • More than one answer may be applicable • Answers do not corresponded to the question • Meaning can be misconstrued • Voting, looking for variation of responses, the larger the scales are, more variation we can tease out of it • Other problems are people that do not provide clear instructions on how they should complete the survey • Which categories are each people in? • Idea of symmetry -> when answers don't correspond to questions • Can you understand the value of something and do it? Question formats: • Single indicators • Multiple indicators o Likert Scales • 1-5, usually a neutral point o Thurstone Scales • 11 points with neutral in middle o Guttman Scales • Ladder -> where you place yourself on the social scale o Semantic differential scales • Two dimensions of a concept-> here we have things like an extrovert and on the other end you have introvert • Series of scale where you place yourself in the dimension where you feel closest to • Extrovert may place themselves somewhere -> tease out a dimension of personality • Series of at least 4 elements for each scale that you would make it so that it is not all positive and negative • Force the person to read it o Create a scale of some sort of concept How to ensure good results -Pilot and pre-test questions • Piloting and pre-testing questions o Check that the instrument works o Open questions can generate fixed choice, answer to include in the final study o Use existing questions Coding and inputting questions • On a scale of 0-10 where - is strongly agree and 10 is strongly disagree how much do you agree or disagree with the following statement? The earth simply cannot continue to support population growth at the present rate • How often do you make a special effort to sort glass or tins or plastic or newspapers and so on for recycling? The SPSS Data Editor • Data viewer o Spreadsheet of 'cells' into which you enter data o Columns represent variables o Rows represent cases • Variable Viewer o Displays information about variables o Lets you define and modify variables Defining variables • Open the variable viewer • Click on default variable name (var00001) and type in your variable name (eg. Reasons) • Give the variable a more detailed label • Define value labels (codes) o Unless interval/ratio variable o Define missing values Questionnaire Construc
More Less

Related notes for COMM 2002

Log In


Don't have an account?

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.