The research process, surveys, questionnaires & interviews Communication Research Week 3 Steps in the research process 1) Establish the need for research 2) Define the problem 3) Establish research objectives 4) Determine research design 5) Identify information types and sources 6) Determine methods of accessing data Communication Research Spring 2005 2 Steps in the research process 7) Design data collection forms 8) Determine sample plan and size 9) Collect data 10) Analyse data 11) Prepare final research report Communication Research Spring 2005 3 The research process DETAILS OF STUDY Purpose of the study Types of investigation Exploration Establishing: Causal relationships Correlations Group differences, ranks etc Description Hypothesis Extent of researcher interference Minimal: Studying events as they normally occur MEASUREMENT Study setting Contrived Non contrived Manipulation and/or control and/or simulation Measurement & measures Operational defn Items (measure) Scaling Categorising Coding Data-analysis Unit of analysis (pop to be studied) Individuals Dyads Groups Organisations Machines Sampling design Time Horizon Probability/ non probability One-shot (crosssectional) Sample size (n) Longitudinal 1.Feel Data-collection method for data Observation 2.Goodness Interview of data Questionnaire 3. Hypothesis Physical measurement testing Unobtrusive Communication Research Spring 2005 4 The Classical Ideal of Science Induction Theories Deduction Generalization Hypotheses Method & measurement Operationalization Observation Communication Research Spring 2005 5 Deductive vs Inductive Reasoning DEDUCTIVE Argues from the general to the particular Eg you observe that all deciduous trees lose their leaves … You therefore reason that your bare tree is deciduous INDUCTIVE Argues from the particular to the general Eg if you burn one finger on a hot stove … You therefore reason that you could burn all of them Communication Research Spring 2005 6 Preparing an hypothesis or research question An hypothesis is a ‘reasonable scientific proposal’ or a ‘statement of expected results’ It is not a statement of fact but a declarative statement which tells the reader what you are going to do NOT how you plan to do it An hypothesis is common to scientific research methodologies while qualitative research more commonly uses research questions to focus Communication Research Spring 2005 7 Writing a hypothesis A well written hypothesis Is stated in declarative form Posits a relationship between variables Reflects a theory or body of literature upon which it is based Is brief and to the point Is testable Communication Research Spring 2005 8 Developing an hypothesis An hypothesis should be developed from a wellresearched body of knowledge which is both logical and feasible eg Inland waterways are becoming polluted through fuel discharge from petrol-driven powerboats Powerboats are noisy thus creating auditory pollution Alternatives to petrol engines exist Electric (battery-powered) boats will decrease pollution on inland waterways Communication Research Spring 2005 9 Ideas, questions & hypotheses Research Interest or Ideas Research Problem or Questions Hypothesis Open classroom and academic success What is the effect of open vs traditional classrooms on reading level? Children taught reading in open classroom settings will read at a higher grade level than children taught reading in a traditional setting. Test-taking skills and grades Will students who how to “take” a test improve their scores? Students who receive training in the “Here Today – Gone Tomorrow” method will score higher on the SATs than students who do not receive the training. Television and consumer behaviour How does watching television affect buying behaviour of adolescents? Adolescent boys buy more of the products advertised on television than do adolescent girls. Drug abuse and child abuse Is drug abuse related to child abuse? There is a positive relationship between drug abuse among adults and their physical and psychological abuse as children Adult care How have many adults adjusted to the responsibility of caring for their aged parents? The number of children who are caring for their parents in the child’s own home has increased over the past ten years. Communication Research Spring 2005 10 Rats & scientific research Communication Research Spring 2005 11 Problems with asking questions as evidence (from Vance Packard (1956) The Hidden Persuaders) 1. You can’t assume that people know what they want In a survey of male drinkers the men expressed a strong preference for a ‘nice dry beer’. When they were then asked how a beer could be dry, they were stumped.Those able to offer any answers at all revealed widely different notions. Communication Research Spring 2005 12 Problems with asking questions as evidence (from Vance Packard (1956) The Hidden Persuaders) 2. You can’t assume people will tell you the truth about their wants and dislikes even they know them. Psychologists at the McCann-Erikson ad agency asked a sampling of people why they didn’t buy one client’s product – kippered herring.The main reason the people gave under direct questioning was that they just didn’t like the taste of kippers. More persistent probing however uncovered the fact that 40% of the people who said they didn’t like kippers had never in their entire lives tasted kippers. Communication Research Spring 2005 13 Problems with asking questions as evidence (from Vance Packard (1956) The Hidden Persuaders) 3. It is dangerous to assume that people can be trusted to behave in a rational way. A test was designed to establish the influence of the package on the product. It gave housewives three different boxes filled with detergent and requested that they try them all out for a few weeks and then report which was the best for delicate clothing (Note: actually only the boxes were different, the detergents were identical – one box was predominantly yellow; the second was blue and the third was blue with splashes of yellow). Result: the detergent in the yellow box was too strong, in the blue box it left the clothes dirty and in the blue and yellow it was ‘fine’ and ‘wonderful’. Communication Research Spring 2005 14 Survey research Purpose is to determine the current status of a population with respect to one or more variables Can be qualitative or quantitative, depending on data sought Communication Research Spring 2005 15 Advantages of surveys Inexpensive Can obtain current information Enable the researcher to obtain a great deal of information at one time Provide quantitative or numerical data Very common and so some of the info you seek may have already been gathered eg ABS Communication Research Spring 2005 16 Problems with surveys People often don’t tell the truth, especially about personal matters People make mistakes about what they’ve done Obtaining representative samples is frequently difficult People often refuse to participate Relatively small percentages of people answer and return questionnaires Writing good survey questions is difficult Communication Research Spring 2005 17 Self-administered questionnaires ADVANTAGES Inexpensive No interviewer bias to worry about You can ask about very personal matters You can ask complex, detailed questions DISADVANTAGES People may misinterpret questions Low response rates the norm You don’t know who actually filled out the questionnaire Sampling errors frequent Communication Research Spring 2005 18 Stages of a survey 1. Define your information needs what information? purpose of information? who knows the answers? practical issues eg time 2. Carry out background research – what work has been done before in this area? 3. Choose a survey technique questionnaires diaries individual interviews group discussions Communication Research Spring 2005 19 Stages of a survey 4. Define and test your methodology 5. Administer your survey – consider any problems Open or closed questions? Pre-test questions Ordering of questions Trial questioning techniques By mail/email Is target group representative? Were all questions answered satisfactorily? 6. Analyse your results Did enough people reply? Is target group representative? Were all questions answered satisfactorily? Communication Research Spring 2005 20 Stages of a survey 7. Present your findings Usually in report format Outline parameters of survey critically Do your results prove or imply results? Acknowledge limitations eg time, sample size, demographics of sample etc Communication Research Spring 2005 21 Problems with Survey Research Inadequate response Including unrelated items on questionnaire Poorly worded items Complex items Leading questions Assuming facts not necessarily in evidence Analysing open-ended questions Communication Research Spring 2005 22 Problems with Survey Research People often don’t tell the truth about themselves People make mistakes even if they are trying to tell the truth eg ratings books Obtaining representative samples is difficult Relatively small percentage of people answer and return questionnaires Communication Research Spring 2005 23 Survey & Questionnaire design When designing a survey or questionnaire you need to consider the type of question which will give you the most accurate data There are five main types of questions 1. Close-ended questions which use yes/no responses Q. I have good communications with my supervisor A. Yes No Communication Research Spring 2005 24 Survey & Questionnaire design 2. Open-ended questions which allow the respondents to give an unlimited answer Q. What problems are you having with your supervisor? Communication Research Spring 2005 25 Survey & Questionnaire design 3. Checklist which present a list of items where participants are asked to check those items that apply to their particular situation. Q. Please check the following types of communications that you have with your supervisor. Informal meetings After hours discussions Formal meetings Telephone Written reports Social gatherings Emails Committee meetings Communication Research Spring 2005 26 Survey & Questionnaire design 4. Multiple-choice questions which offer several choices and the respondent is asked to select the most correct one. Ensure the choices presented cover all the possible options. Q. How often do you purchase items from the company vending machines? a) Once a day b) 2-3 times a day c) 3-5 times a day d) 6 or more times a day Communication Research Spring 2005 27 Survey & Questionnaire design 5. Ranking scales which require the participants to rank order a list of items. Q. Of the following list of five types of communication that you might have with your supervisor, rank from 1 (most important) to 5 (least important). — Formal meetings — Informal conversations — Written reports — Letters or emails — Telephone discussions Communication Research Spring 2005 28 Survey & Questionnaire design 6. Likert scales which usually measure attitude toward a concept or idea. It allows the respondent to indicate the degree of agreement usually on a 5 or 7 point scale. Q. Please indicate the degree to which you agree or disagree with this statement. “The 1995 Ford Falcon is a substantial improvement on the 1994 model.” Strongly agree Agree Neither agree Disagree Strongly nor disagree Disagree 5 4 3 Communication Research Spring 2005 2 1 29 Survey & Questionnaire design 7. Semantic differential scales measures attitudes by displaying pairs of opposite terms and asking respondents to check which term best describes their feeling toward the concept or topic. Q. Place an X in the space between the two terms that best describes how you see the XYZ Corporation Office. XYZ CORPORATION DISTRICT OFFICE [the topic or entity being evaluated] Pleasant - - - - - - - - - - - - -- - - - - Unpleasant Efficient - - - - - - - - - - - - -- - - - - Inefficient Not helpful - - - - - - - - - - - - -- - - Helpful Professional - - - - - - - - - - - - -- - Unprofessional Insensitive - - - - - - - - - - - - -- - - Sensitive Friendly - - - - - - - - - - - - -- - - - - Unfriendly Slow - - - - - - - - - - - - -- - - - - Fast Rigid - - - - - - - - - - - - -- - - - - Flexible Communication Research Spring 2005 30 Other types of qualitative research – depth interviews one-on-one interview which is relatively unstructured with a subject by a trained interviewer. The direction of the interview is guided by the responses to the questions Communication Research Spring 2005 31 Other types of qualitative research – personal interviews ADVANTAGES Interviewer can explain Qs in detail Interviewer can use a variety of data collection methods Interviewer can spend a lot of time with respondents You know who is answering the questions A higher likelihood of achieving the desired response rate Not intimidating DISADVANTAGES Can be intrusive (too personal) Time-consuming and expensive Hard to find people in sample at times People are reluctant to answer some questions Needs well-trained interviewers Possible language barriers Communication Research Spring 2005 32 Other types of qualitative research – focus groups Often used in market research A group is asked a series of structured questions and guided through structured sessions Designed to probe their attitudes and feelings about a range of issues Communication Research Spring 2005 33 Other types of qualitative research – projection techniques Such as word association tests, sentence and story completion, cartoon tests, consumer drawings, photo sorts These techniques belong to the field of clinical psychology They are designed to probe and penetrate a person’s defense mechanisms and allow true feelings to emerge Communication Research Spring 2005 34 Communication Research Spring 2005 35