CHAPTER ie COMMUNICATIONS RESEARCH METHODS “... digital media collectively provide searchable access to a wealth of experiences and insights, the quantity and diversity of which seems likely to increase substantially.” — Jonathan M. Levitt CconrnsB070R 70 THs JOURNAL OF BlCHRAL INFORMATION PCN ud eg46 cuapren} Communications research methods GRP RESEARCH Gathering and using information to clarity an issue and solve a problem, [hts tse research inthe ommanteions fd has enploded Before his time, it was often considered an unaffordable luxury. Now we seer to abound in data, so much so that we sometimes drown in it. The key, now, is to sort through and synthesize data into usable information to help us make wise decisions. Not only do most successful organizations now do research, but there are @ plethora of specialized consultants and research firms which have taken market, social media, environmental, communications and organizational research to ev. els of sophistication never before dreamed, Whereas in the past, communications practitioners found themselves begging for a pittance to find out what their pub- lics thought, they now have executives whose first question is, “What does the research say?” {As a result, our challenge is no longer to convince practitioners to do research. At this point in your education or career, you have learned the value of research and measurement and how to do it or how to buy it, Further, there are now doz- ens of texts and handbooks a5 well as online resources for conducting research. ‘The challenge is to provide the basic framework for thinking about and organizing research and analysis, and then, in the next chapter, t0 appiy it in the strategic: planning process. The role of research in communication Research is only as good as its application to the problem-solving process. To be research-oriented means gathering and basing decisions on information as part of your daly routine, To be effective in communicating with an organization's publics, ‘we must be constantly listening and scanning the environment for information. We should establish good communication channels s0 information is constantly flow. Ing tous — resulting in adjustments and refinements of our efforts as plans proceed. Research helps us to Save time and money. Undersiand our publies Make sound decisions. Avoid mistakes. Discover new ideas. Identify potential publics. Identify communication channels. Justify plans. Connect with communities As professional communicators we should be wary of “gut reactions,” knowing ‘what we know about how peaple perceive and misperceive. Always test the infor mation that leads to conelusions, and especially to key decisions. The next chapter provides checks afthe information you need to meet the various challenges of an ‘organization and to pian strategically to seize opportunities. In this chapter, we iden- tify some of the best sources of taformation and the methodologies used to obtain it.CcHAPTER 3 Communications research methods 47 Research methods and the diversity of tools Research methods are often categorized as formal and informal, quantitative and qualitative, and primary and secondary. Nevertheless, these categorizations are not poallel, For exemple, formal research is not necessarily quantitative research, nor is it always primary research. A fewr definitions regarding research will help to avoid confusion. Formal and informal research, Formal teseerch implies @ structured study, It is governed by rules of research that include previously identifying what you hope to learn, how and from whom, Because it follows universal rules of research, the findings are more accurate and reliable. Informal research. is less structured and more exploratory. It does not follow specific rules, Nevertheless, it often provides valuable insight to lead us in directions of more formal discovery. Quantitative and qualitative research. Quantitative research gathers statistical data for analysis. Qualitative research is focused on individuals and groups not statistically representative of 2 given population. While qualitative re search may be supported by some statistical data analysis, itis not governed by laws of probability. It may, however, be governed by rules of research, Focus groups, for example, are a qualitative tool. They may be informal discussions, but are more often formal research yielding important insights into a publi’s perceptions, attitudes and motivations. Now that focus group methodology has actually become a dominant method of market research, the rules of research governing this methodology are meticulously followed to ensure accurate insights. Although the method is classified as formal re- search because it follows rules and structure, itis stil typicelly a qualitative approach that yields in-depth understanding but no statistical data Primary and secondary research. Primary research implies gathering the infor- mation firsthand fora specifically identified purpose. It doesn't necessarily refer to survey research. Personal interviews as well as mail and telephone analysis also yield primary information, Primary research is research that ‘you implement yourself or contract out fora particular purpose. Secondary research is primary research data originally collected for a different pur- pose that is now being drawn upon for a new use. Typically itis cheaper and faster to use secondary research, In this era of omnipresent data, you should exhaust secondary sources before embarking on any costly primary research efforts, Given these definitions, the research tools become more difficult to categorize Focus groups may be formal or informal. They are typically qualitative, but ifenough soups are conducted, some quantitative data analysis may be done on the results, Results are primary research when you organize and condutt them for the immedi ate purpose, but reviewing transcripts and analyses of focus groups conducted for ‘other purposes is secondary research that may shed light on the problem you are ‘ying to solve. PRIMARY RESEARCH Firsthand information gathered specifically for your current purpose, INFORMAL, RESEARCH Less-structured exploratory information gathering, ee tat pronae dene significance, SECONDARY Ry RESEARCH Information previously assimilated for other purposes that can be adapted for your needs,48 cnaeten 3 Canmunicsions esearch mesos Similarly, personal interviews may be informal and qualitative resesech. They ray be one of your “listening” techniques. Os, given more structure and an 3p- propriate design, they may be formal and quantitative, allowing statistical analysis with a high degree of confidence. They would be a primary research tool if con- ducted for the project at hand, yet may be useful as secondary data in subsequent, programs. Whether the research you do is formal or Informal, quantitative or qualitative, primary or secondary, depends largely on what you need and how you structure it. You should determine your purpose (what you are hoping to accomplish with the research) and what you are trying to find out from whom, before you decide on the best tools to use and how to structure the effort Secondary research Organizational research ‘The frst place to begin in gathering information is within the organization itself Many kinds of research tools are available to help gather and assess the information available to you, A communication audit examines all of the organization's commu- nication to see if t supports the organization's mission and message. Environmental scanning within an organization monitors the mood and feelings that exist among the workers, customers, investors, suppliers and many other publies of the organiza tion. Online, mail and telephone analysis helps you track whit issues cause concern among your publies. Social inedia analytics help to monitor your reach and relation ships. Certainly customer service and complaint sites help you track opinion trends, and potential problems. Important background information about your company or your client is found in the publications andl websites of the organization. Employee publications, blogs, digital communications, annual reports, brochures and marketing materials, policy and procedates manuals, organizational charts, sles and accounting records, histo: ries and any other material available from the organization, either in hard copy or electronically, ean be valuable information, Keep in mind that such material usually possesses an inherent bias, and you need to loak outside the organization as well as inside to make sure you have the complete picture. Organizations do not often open thelr elosets to display the skeletons through their ‘own printed and electronic material. You will get rich information about the organization theough its ‘material, ut you will not usually get the bad news, And, not knowing the bad news may sabotage your communication efforts, While the organization itself js the place to start your research, it should never be your only source. ‘he organization may also have data from past surveys or research. You may need some primary research to determine the mood and opinion of ‘employees, Most organizations would benefit by taking a searching look inside before focusing re search efforts externally,Ccomnrren 3 Communications research methods Internet and library research {nformation technology and the computer revolution have given us acces to incted ible resources for research. Information that would take weeks or months to find fom original sources is now readily avaiable a our fingertips. The communication professional of today and of the future must understand how to get good data from the Internet to compete in this new environment, ‘Online research has, in most cases, replaced @ personal visit to the library. It gives the researcher access to the collections held by thousands of libraries and or- ganizations and to databases full of information and references, Nevertheless, much is availadle in library documents that may be difficult to find or expensive to secure elsewhere. Most of us underestimate the value of the data available in our local col- lege and public libraries. And unless you actually take the time to investigate, to talk to a resource librarian or to just explore the collection, you will not appreciate the vast amount of information at your fingertips Remember the latest census? It's available on the Internet. In that census, some people received a more in-depth questionnaire than the rest of us, and the psycho- graphic data are accessible as well. Further, the census is continually updated with Interim studies. Also accessible isa host of government documents and studies pub: lished every year, along with many private research studies. You can find national and local newspapers and magezines that date back years, sometimes to the begin. ning of publication. Often the results of opinion polls can be found and rich eco- nomuic data on local, state, national and international markets. Some universities have separate libraries for ther business schools. In that case, the business library probably contains detailed market analysis and other such valu- able information. While most of this information may be available electronically, there are volumes of information that can only be accessed by visiting a library in person. The more current the information, the more likely you will be able to get it via the Internet, Nevertheless, there may be significant risks in ignoring the deeper background that older documents contain. Organizational information is also readily available on the Internet, And in- creasingly, you will find many independent websites, blogs and other Internet sites that contain valuable information, Be careful of the source of the informa- tion, and seek secondary confirmation when possible. The information may be credible, such as that from an industry analyst with a professional responsibility to provide such information, or it may be a site constructed by a disgruntled cus: tomer or employee containing extremely biased or inaccurate information, rumor and innuendo. External organizations ‘Our tax dollars support local, state and national government offices that have a charge to operate in the public interest. Providing public information is often an in tegral part of that responsibility. Much of the informnation is now available over the Internet, but some ofthe valuable information you seek may only be available upon request, or by digging through studies and papers. Sometimes the bureaucracy can be difficult, and getting information can take weeks or even months, Nevertheless, the information available is often critical. “Most cities and states have economic development offices of some kind that cal lect invaluable information on industries and markets. State, and sometimes local, governments have information on population, wages, education, unemployment, 4950 cHarreR3 Commamicaons rscrch methads Q._, DEMOGRAPHIC oP pata Information used to segment publics according to tangible characteristics such 25 age, gender and socioeconomic status health and just about everything else you can imagine, Environmental data is read: ily available from area chambers of commerce and travel councils. Be persistent; the information you want may be part of something else. You will have to do most of the searching, so start specific but be ready to broaden your search until you find documents and reports that will provide the information you need. You have to ask broad and searching questions to get to the right documents and studies. For example, asking a clerk for any studies on how much college students spend on e-books might not get you the information you want, But studies of online se or income and expenditures of 18- to 24-year-olds or the costs ofa college education would all contain information on e-book expenditures. Inherent in the missions of associations, advocacy groups or professional sock tes is gathering and disseminating information. One of the most valuable benefits ‘of membership may be access to the research they gather. You may be able to access the information you want from them through their publications or resource librar- ies. You may have to pay a search or use fee to access the material. In some cases, you may need to get the materia through an association member. But the data available is generally very rich, current and valuable. ‘A word of caution: When you receive data from these kinds of organizations, especially from activist groups, check the souzces and methodologies used. Be aware that any information published by an imerest group of any kind will be inherently biased to some degree, Make sure you understand and allow for that bias, and seek confirming and/or disputing information from other sources, or carefully examine the research methodology used and adjust for distortion, Media research “Media research straddles the line between secondary and primary research. Depend- ingon the purpose, itcould be either. Nevertheless, itis erucial in today's media envi ronment. A number of publishing houses produce Internet-accessible media guides and services that provide current and valuable information about media throughout the nation by category: newspapers, magazines, radio stations, television stations, cable stations and so on. Media services such as PR Newswire, CisionPoint, which is, based on Bacon's media database, and BurrellsLuce media monitoring track editors and reporters by assignment, how to submit pieces aa what is typicelly accepted. The guides also indicate readership, wiewership or listenership and will sometimes provide additional demographic information that stay be of help in profiling key publics, They at least provide a way to contact the media organization to request more detailed information. Most media organizations can provide detailed viewer, listener or reader profiles because they sell advertising, And, advertisers want to know who they're buying access to. ‘Media and Internet analysis and clips are @ critieal part of communications research and evaluation, Whether in-house or a contracted service, tracking such coverage is essential, Nevertheless, because of the time it takes to be thorough, you will usually get more comprehensive and cost-effective clipping and analysis, if you contract a clipping service. Clipping services may do as little as simply clip anything (print, bro2dcast and other electronic media) that mentions the com. pany oF an issue of interest to the company, At the other end of the spectrum, they may engage io extensive analysis and evaluate the positive or negative int pact of the pieces that discuss the organization and its competitors or any of thecurren 3 Communications rscerch metho — 51 issues faced by the industry. You can specify the level of service you want and pay accordingly Monitoring the Internet is particularly important to organizations today. And the methodology has become more sophisticated as tracking tools have been devel- ‘oped and refined. Most organizations have given up teying to monitor al the online conversations, but a selective approach to online listening related to Issues of con- cer to yout organization isimperative. Primary research Focus groups Focus group research has become an important and reliable source of data to under stand our publics. & focus group is a moderator-led discussion with four to 14 per ticipants. The moderator asks open-ended questions to garner qualitative responses oon attitudes and behavior. The moderator must be careful not to bias the discussion by injecting personal opinion or information into the group. He/she should encour age participation from all members ofthe group and probe for in-depth understand: ing, The moderator must also ereatean atmosphere of openness, honesty, safety and confidentiality in order to engender free and open discussion. With the permission and knowledge of the participants, he session is usually recorded (audio or video), and the discussion transcribed for further evaluation and data tabulation. Focus group research is generaly easier to conduct than survey cesearch and provides rapid results along with depth of opinion and attitudes within the group. The discussion nature of focts groups shapes, refines and crystallizes opinions and attitudes (see Figure 3), Further, while not always less expensive than other kinds of research, it is often more cost-effective. Focus groups used to be conducted in communications ané marketing research primarily as discus sion forums for advisory committees or idea panels to supplement quantitative research, The information was often used as a precursor to survey research to assist in developing a questionnaire that adequately probed attitudes and opin ions. In today’s research-oriented marketplace, many practitioners recogni that, while survey research is becoming less credible as an accurate representa tion of publies, focus groups provide the kind of information needed to immedi- ately address and resolve problems. {As discussed in the previous chapter, people do not behave in isolation, The dis- cussion and refinement of opinions and attitudes which occur in focus groups often provide problem-solving behavioral information that surveys cannot. The example cited in the previous chapter of the introduction of new Coke is a case in point. Both research techniques were used, yielding opposite results. Yet because of the reputation for validity of survey research over focus-group responses, the company chose to rely on predictions of behavior based on attitudes expressed! in isolation ‘They should have more carefully considered the g:oup behavior that emerged from the focus groups. Further, focus group research can demonstrate the process of opinion forma: tion, While not representative, the group isa social microcosm of a larger public. The analysis of how attitudes and opinions change based on the flow of the dis FOCUS GROUP RESEARCH Moderatorled discussions with fewer than 15 participants providing in-depth information on attitudes and behaviors. ° aotS52 CHAPTER 3 Communications research methods gy Figure 3.1 © “Uses and abuses of focus group research Uses Abuses Immediate results. Weak moderator. Comfort in numbers. Dominant moderator. Flexible and response-orlented. ‘Not homogenous. Gauge of group behavior. Too few groups. Issues oxplored and opinions crystallized. Generalizing to a population. ‘Sensitive issues addressed. Attitudes of activists Included. {ssues and jargon identified.‘cuAarTeR 3 Commuviaion research meade 53 cussion can help us know what information people need to make sound decisions and what appeals will be most effec- tive in the larger arena. Innovations in focus group research, now allow quantification of results if certain conditions are met. Conducting large numbers of groups and employing some content analysis techniques can make the data statist cally reliable. The Internet provides an interesting resource for tea ditional focus group research. Scheduled online chats can produce similar results without the geographic restrictions. Nevertheless, care must be taken that participants are invited and known, Otherwise, the data may not be useful for the re searcher’s purpose. Face-to-face group discussion is still preferable but can be simu lated using tools such as Skype. © LD Potshunesockcom Copy and product testing One classic use of focus groups is for copy and/or product testing, But this is not the only method by which to test. Copy testing simply selects individuals within your target publies and requests their review of copy, whether survey copy or communication copy (brochures, advertising and the like). In product testing, individuals are asked to examine and use a product and provide feedback on everything from packaging and sales methods to product quality. Product tests may be done individually by personal interview, by mail, in focus groups or online Honest responses in copy snd product testing help avoid costly mistakes. Survey research instruments should always be tested before being implemented. ‘Testing copy helps ensure that the messages are coming across in such @ way as to produce the desired result. Marketers test promotional campaigns or products, in areas representative of the overall market. Sometimes, two or three different versions of a product or 2 campaign will be tested in similar areas to determine which will be the most effective to roll out nationwide. Copy and product test ing is one of the most valuable kinds of research available to the practitioner. Its greatest value lies in its ability to prevent mistakes — saving money, effort and time. Psychographic studies Values and Lifestyles Segmenting, developed by SRI International in 1978, is re search methodology that classifies publics not just by demographies, but also by psychographics or atitudes, beliefs and lifestyles. Found to be far more effective in segmenting publics than demographies alone, psychographie studies help us 10 know what motivates individuals within a particular public. The VALS categories — achievers, survivors, sustainers, belongers and so on — have been used extensively in advertising and marketing to segment and tailor messages to specific target pub- lies, They provide the same valuable segmentation for communication with all the ‘organization's publics. Communicators should know the VALS eategories, both the original and the more recently revised segmentations, and understand the motiva tions tied to the differences in attitedes and lifestyles.$4 cunnreny Communication rssh methods Cag PSYCHOGRAPHIC oP Dara Information used to segment publics according to values, attitudes and lifestyles. Other market research firms have subsequently developed similar categoriza- tions. In the 19905, Claritas developed PRIZM, segmenting American consumers into 14 groups with 66 demographically and behaviorally distinct types like “Cosmo- politans” and “Kids & Cul-de-Sacs.” PRIZM was subsequently purchased by Nielsen, the company known for rating what people watch on television. Oftentimes, local media and other similar organizations will have segmented and profiled their own target audiences using a combination of demographic and psychographie data, Whereas some will be unwilling to disclose the information, ‘which is quite costly to compile, others may be persuaded to share the data, especially ifthe request comes from a nonprofit organization or is for a charitable purpose. Another valuable tool for understanding key publics is Values in Strategy Assess ment, a process developed by Wirthlin Worldwide and purchased uy Harris Interac- tive, which has now also joined forces with Nielsen, This tools premise is that values are the fundamental determinant of an individuals behavior and decisions, Under standing the fandamental values of a public provides the strategy to motivate action. Social media analytics ‘A new reallty of today’s communication environment is pervasive social media. With the range of networks available today, organizations must carefully choose those with which they will engage ~ those most appropriate for building relation- ships with key publics. Facebook was perhaps the first to became an essential pres ence for the organization. Now Twitter, Pinterest, Instagram, Snapchat and a host of ather social media networks have become channels through which we interact with our publiss, Since the effort to build and manage a social media presence is so costly and time-consuming, itis critical that organizations employ research methods to ensure the effectiveness of their social media efforts in delivering the messages of the organization, building relationships with key publics and providing the interse tions key publics desire Quite simply, social media analytics are the tools used to measure, analyze and interpret the interactions and relationships with our key publics. I, in large part's essentially “online listening” to help us better meet the needs of publics. The East rman Kodak Company used online listening several years ago to determine what features consumers wanted in a digital camera. They discovered through theit re search that no existing camera offered the combination of features consumers were bbuzzing about on Facebook. Kodak engineers quickly went to work and within a few manths were able to bring a new product to market that contained all of the features people wanted most at an affordable price. The new camera sold extremely well, but unfortunately wasn’t enough to save the struggling company from bank ruptey in 2072 ‘This kind of measurement is so important that many of the social media net ‘works, like Facebook and Google, have imbedded their own monitoring and ana Iytical tools to track sociat media efforts. In addition, a host of external tools — some fee, some for a fee — integrate measurement of all social media efforts, It is essen tial, nevertheless, that you are careful about what you are measuring. Measuring the number of “ikes" or “Friends” or “followers” doesn't measure whether publics are actually retaining and acting on your message. Don't confuse coverage with success. Remember that behavior — or action — is the final evaluation. Itdoesa’t matter how ‘many “followers” you have ifthey don't actually perform the action that will satisfy ‘yout objectives.cuarren 3 Communications rsuch mattods 35 Survey research and opinion sampling ‘A popular quantitative research method is survey research, although its credibility hhas declined somewhat in recent years because of the difficulty in securing a truly random sample. Several events have affected the ability of researchers to secure truly random and representative samples, Mail surveys have always been extremely unreliable, not only because of low response rates but also because of skewing. Only certain kinds of people will take the time to respond to a mail survey, making it any- thing but representative. ‘Telephone surveys have also declined in eredibility for the same reason. And the “do not call” registry has mage telephone surveying more difficult than ever. Although telephone surveys are not prohibited by the registry, the very existence of “do not call” list seems to have given people the courage to refuse calls they would have previously endured. "Nevertheless, survey research has been 9 popular research technique in com munications and will probably continue to be in some form or another. For exam ple, researchers may find the personal drop-off method to still be effective because of the personal contact involved in dropping off and/or picking up a survey. Its not as easy to turn down someone face to face as it isto say “no” over the telephone. Online programs like Qualtries make surveying easier but still face the challenge of securing representative samples. Professional research firms have developed other techniques that ean also improve response rates. Survey research is a difficult and exacting approach. It requires meticulous at tention to detail at every step ofthe process: questionnaire design, sample selection, survey implementation, data processing and data analysis. A mistake or misjudg- ‘ment at any point will skew the results, often without the researcher knowing the data is skewed, To be valid and reliable, survey research must follow strict rules of research. ‘The idea behind survey research is to take a sample from a population, or universe we follow good statistical procedures, we should be able to make that sample relatively representative of the universe, although we can never be absolutely sure of our accuracy unless we survey every individual in the population (a census). The total number of individuals surveyed in the population and the way they are selected ‘will determine how accurately the results reflect the universe Statistical research on very critical issues, or in close political campaigns, needs to have a high level of confidence and a low sampling error, also known as margin of error. The confidence level reflects the researcher's percentage of certainty that the results would be the same (within the margin of error) upon replication of the survey. The margin of error reflects the percentage points thatthe semple results, on any given question, may vary from the population asa whole, Increasing the sample size increases the confidence level and decreases the margin of error. The only way to be 100 percent confident and eliminate the margin of error would be to survey the entire population or take a census. Research regarding an organization's publics generally requires at least a 95 percent confidence level and @ margin of error of 5 percent or less. Further, the more important or controversial the issue, the greater the need for a lower margin of erro. “There are two basic kinds of survey sampling: probability and nonprobabil- ity. Probability sampling is scientifically random; every individual in the popula tion has an equal chance of being selected, Nonprobability samples survey who- ever is available, for example, intercepting students during the lunch hour as they SAMPLING ERROR GL, Meesured os moran of tron nda the oossble percentage ‘eration of sop dataromthe whole oonulton CONFIDENCE LEVEL The percentage of certainty that the results of a survey would be the same if replicated.56 CCHAFTER 3 Connumicatanseierch nerds TIPS FROM THE PROS Beyond sampling error Dee Allsop, Ph.D., CEO of Heart+ Mind Strategies and former president of Harris Interactive Solutions Groups — producers of the Harris Poll — tips you off on understanding key factors in survey accuracy. With the current emphasis on “big data® for critical decision-making, survey and sampling science are all about accurately measuring a small number of people to understand the attitudes and predict the behay- jor of an entire group. Accuracy in survey research is the direct result of ‘tentifying. eliminating or correcting for aif types of error or bias that can enter into research. While “sampling error” is the most visible indicator of survey accuracy reported in the media, your success depends on under- standing both sampling and nonsampling errors. SAMPLING ERROR ‘Sampling error is the statistical estimate of variation that will occur when using a small number of people (a sample) to approximate a result from a much larger population. One way to think of sampling er- ror is the range of results that would occur if you were to repeat your survey hundreds of times. For example, many surveys report a sampling erfor of plus or minus 3 percent at the 95 percent confidence interval. ‘This simply means that if you were to repeat this study 100 tires, 95 percent of the time your results would fall within a range of 6 percantage points. NONSAMPLING ERROR ‘Sampling error is inherent in ail surveys. We understand it well and can predict it, Other sources of error are more elusive, yet far more important to accuracy and utility of survey results, ‘Sample frame. Make sure the sample was taken from the group of people that matter. For example, if the survey is about who will be elected pre dent, the sample should come from people who can actually vote for presiclent: registered voters. Nonresponse error, Several factors influence a respondent's likelihood to respond (e.g., mode effects, interviewer effects and sensitivity effects), Rather than sampling error, | would much prefer to know the response rate for a survey. & law response rate indicates the survey Is less likely to reflect the population in question.CHAPTER 3 Commamications reeurch methods 57 Construct validity. Questions can be worded or constructed to gener- ate just about any result desired. Always read the questions that were asked before interpreting the results. Biased wording will produce biased resutts, Institutional reputation. Always check to see who conducted the survey ‘and who paid for it. Reputable companies invest significant time and re- sources to produce objective findings and eliminate nonsampling errors. ‘Take a lesson from marketing and survey research companies. Instead of worrying about “sampling error,” focus on understanding how to reduce nonsampling error through goad survey design, questionnaire construc- tion, interviewing execution and data processing and correction, Copvigh ¢ Kendo Hunt Ua ComoRTY enter the student center or interviewing people at a grocery store on Saturday afternoon. ‘There are also two kinds of errors: sampling and nonsampling, Sampling error is inherent because we are surveying a portion of the population and not the entire population, It is the percentage of possible variance in the sample's answers from the population, We decrease sampling error by increasing sample size; the closer ‘we get to surveying the whole population, the smaller the sampling error. Sampling error is unavoidable in survey research (unless we take a census) and is measured as margin of error Nonsampling error is all other types of errors introduced into the process Misvakes made in questionnaire consteuction, saraple selection, survey implement tion, data entry and tabulation are all nonsampling errors. Great care must be taken in selecting the sample as well as in designing and implementing the questionnaire. Question-and-answer categories must be designed to avoid the introduction of bias so that answers accurately reflect the information the researcher needs. Bias can also be introduced in the implementation of the survey. Ensuring strict confidentiality of responses can lessen courtesy bias. Training interviewers to ask ‘questions without injecting value inflection or personal comments, explanation or other dias is also critical. Further, great care must be taken in coding the surveys and entering the data. Data processing converts the observations ané responses into statistics for interpretation. Data analysis manipulates the data to make logical in ferences. For the inferences to be reliable, the data must be accurately entered and. processed. Finally, the inferences made must be fully supportable by the dataset. few years au, 2 ballot measure in Utah proposed a ligt ral transportation system to be fand- ced by a small tax increase. When the measure was soundly defeated, many analysts concluded it was a vote against light rail. In reality, it was more likely a vote against the fanding method, not the light rail system itself. When we deal with statistics and NONSAMPLING O55, ERROR ~ Mistakes made in designing and implementing a questionnaire that may include efinitional aifferences, misunderstandings and misrepresentations as well as coding errors and/or a failure to represent all populationsSB cuarren s Communications estarch merhoas Sap purposive SAMPLING Identifying and surveying opinion leaders to determine attitudes and behaviors, CRP STRATIFIED SAMPLING Selecting the sample to ‘ensure proportionate representation of segments within the ° (BP panes stupy Respondents who hove repestedly 0 track opinion ar tte make inferences from data, we must be very cateful thet the data support the eon clusions. Otherwise, we have established « faulty foundation far decision-making, In addition to the types of survey research discussed above, other variations have specific purposes and benefits. The following short cescriptions will provide 4 basis upon which tos westigate the techniques for any given research situation Purposive sampling, Based on Katz avd Lazarsield’s two-step flow theory, purposive sampling identifies and surveys opinion leaders to determine attitudes and behaviors, The researcher must devise a procedure that se- lects the target publics influentials (or eauses them to self select) and then surveys opinion and behavior. It is also helpful to know a litle bit about opinion leaders, such as where they get their information about certain issues. Understanding opinion leaders helps us understand how they will influence others, The Edelman Trust Barometer is an example of purposive sampling Stratified sampling. Truly random sampling should yield a cross section of the population representative of the characteristics within the population Ge, proportionate numbers of women and men and so on). Whenever we skew the randomness of the sample by using techniques that make it easier for us to complete the research, like surveying every “nth” number in the local telephone directory, we risk jeopardizing the representativeness of our sample, Ifabtaining a tray proportionate representation is critical, the research sample should be stratified so that it includes appropriate propor tions of the key segments of the overall popultion. Internet surveys. § growing area of survey research is conducting surveys over the Internet and through social media. While there is inherent bias because of the nature of accessible respondents, ahis data can be extremely valu able i the purpose is consistent with the population semple. Increasingly, ‘organizations use this method to survey their members, employees or cus: tomers. Commerelal firms with access to email lists may further facilitate this method. The low-cost nature, ease of conducting and rapid feedback of Internet surveys have fueled their popularity Personal interviews. Very sensitive issues and research that requires deep probing for attitudes and behaviors are best addressed through personal interviews, The personal interview ensures greater control over the sample and the data, But not only is this method costly, it requires alot of tine and well-trained interviewers, Nevertheless, in certain circumstances, its the only viable method to secure reliable and useful information. Benchmark surveys. This type of survey is simply @ periodic reexamination of| attitudes and opinions within the population, An initial survey is done to set a benchmark against which subsequent survey results are compared, Benchmark surveys are good tools for measuring change as well 25 for eval: usting the success of a program. Panel studies. Sometitnes you will want to study attitudes and opinions on a vat ‘ety of issues over a period of time. Panel studies vet respondents who will be available for follow-up surveys at least once and often several times, ForCHAPTER) Communcavions rescorch methods — 59 example, a newspaper will select individuals from its readership to follow a specific issue or election and respond to queries at specific points in the campaign. Behavioral studies are also sometimes conducted by a panel to as sess whether a change in behavior is temporary and what motivates perma nent change. ‘Omnibus surveys. One of the easiest and least expensive methods of obtaining OMNIBUS SURVEY survey data isto add a few questions to an omnibus survey being prepared An ongoing, open survey by a professional research company. These surveys, sometimes called tay to which a company fon surveys, usually have multiple research clients that share the costs and vt organization may the common demographic data collected. Specific data gathered from the add a few proprietary questions each client adds remains proprietary. The advantages are the cost questions at a saving shared by multiple clients and the currency of the data because the reasonable cost surveys are ongoing and conducted regularly, in addition, the reliability of the data is high as the sample size is usually large. Summary (Our environment has changed. CEOs require “big data” to make decisions, and they. expect your recommendations to be logically based on reliable information. To se- tes cnsantflw of tend of eration you hed to make decor, met thalenges and plan sategicaton, You mst inthe ight combintlonofcontine ‘ous research techniques. The purpose of the research and the kind of information desired drive the selection of methodology. Otherwise your research will be useless because it is inaccurate or because it doesn't provide the information you need to design pesusve camel Exercises 4. Volunteer to conduct focus groups for a nonprofit organization. Subse: quently, design and implement a short survey to gather opinions, attitudes and demographics on some of its key publics. 2, Visi the local library and talk with the reference librarian to discover the range of resoutees, databases, search engines and other references avail- able. Search through some of the material to discover the kinds of informa: tion accessible, not just online, but in the library itself. 3. Look up the most recent U.S. Census and write down at least a dozen cat cegories of information available through this rich resoure. 4. Compile list of “go to” sources to quickly find continually updated infor: ‘mation on issues, publics and communication channels,60. ccuarren 5 Communicains sear serods References and additional readings bebo hepato a aed Delman, CA Waawerh Pubaiag pine Df. enim mrs: One for ended atone, me ‘engagenient, and key relationships. Duthamn, NH: Paine Publishing, ene sicusonte