Feeds:
Posts
Comments

Archive for November, 2009

            Kif Richmann

 

Market researchers spend millions upon millions of dollars every year to discover something that on the surface seems pretty elementary: What are people thinking? But it’s more than just this simple question. It gets looked at from every possible angle.  What are people in California thinking? It’s what are people thinking now? What are people in California thinking now while they drive in their car talking on their cell phone while driving to work? The questions go on and on… and good market researchers know that it’s not just about finding the answers, it more about asking the right questions.  Oh, and asking the right people. 

            Many different types of surveys exist to help market researchers ask the right questions and find the correct answers (and ask the right people).  Two types: cross sectional and longitudinal surveys, are the most common forms used by market researchers.  Cross sectional is the use of many different respondents at the same time. This survey can help you determine what kind of person will like or dislike your product, and is a helpful tool for narrowing down a company’s niche demographic. For example, if data from a cross sectional survey shows that men generally rate a product higher than women, advertisers can focus their efforts towards men. But it goes even deeper than that. If a cross sectional survey shows that Latino men ages 18-35 in general rate a product higher than any other sex/age/race demographic, they know it will be effective to focus on that one market. (Assuming the information from the survey is correct.)

            A longitudinal survey is a survey in time. The survey may be given weekly, monthly, annually, biannually, or even more spaced out. The point is, the survey focuses on how perceptions and opinions change over time.  They often include the exact same people originally surveyed (panel survey) or people of a similar demographic (cohort survey).

            A study headed by University of Wisconsin-Madison researcher Aric Rindfleisch focused on eliminating data error in longitudinal and cross sectional surveys, and detirmingin which method was best for market researchers.  The final verdict was a little split.  Basically longitudinal surveys are best for eliminating data error through a more thorough examination process, but cross sectional surveys were adequate for the job, because they offer researcher a quicker answer.            

            If I were a market researcher, I would probably want the quickest answers, and would therefore use cross sectional.  I would probably invest in longitudinal research, but for my dollars I would want answers that turn results as soon as possible. I would also want to know who to target, and where, and cross sectional gives you the answers needed. While longitudinal research gives you important information (how have opinions; perceptions changed) it seems that cross sectional surveys offer the best answers.

            While not related to market analysis, a poll from CBS shows approval ratings for Barak Obama dipping to 53%.  This is down from an approval rating of almost 70% in April of 2009.  This is a prime example of a longitudinal survey. The data has been gathered every month since January of 2009, and has shown how people perceive the new President.  This is a type of cohort survey, because it surveys a different panel every time, in this case, the demographic is very broad: Americans, eighteen and older.  It would be very interesting to see a longitudinal panel survey about the perception of Obama.  Would the same person have the same opinions later? How about six months from now? The researchers in this case would need to apply careful questionnaire design in order to establish a powerful and useful result.

Read Full Post »

David McNace

Reading Reaction

November 19, 2009

Public relations professionals have to work with polls and try to analyze the results for their clients. Many of these professionals spend countless hours trying to understand, explain and predicting marketing behaviors that will exist in the future based off of polling data. However, the issue of validity has arisen with two concerns being the common method variance and the casual interference. In other terms, the dominant problems revolve around the error due to the using only a single source and the ability to infer causation from observed empirical relations.

In the selection “Cross Sectional Versus Longitudinal Survey Research: Concepts, Findings, and Guidelines” the authors discuss how improving cross sectional surveys would be beneficial to helping the public relations professional and offer longitudinal surveys as a replacement for cross-sectional surveys. The cross-sectional survey method is currently the method of choice, but lends to have a high common method variance and casual interference. To remedy this situation it would be beneficial to also collect longitudinal data. Cross-sectional surveys are completed by a single respondent at a single point in time, where as longitudinal surveys are where data can be collected over multiple periods of time. To produce a reliable survey, researchers have recommended various collection strategies. These strategies include employing multiple respondents, obtaining multiple types of data and gathering data over multiple periods. The authors believe that longitudinal surveys are a better method in collecting information, despite both types of surveys being an adequate means of measuring information. However, many companies stick with cross-sectional surveys because of the high levels of cost and time that are consumed by longitudinal surveys.

The APR reading discussed the pros and cons of polls, which make participation easier and individuals’ participation can remain unknown. The authors also stated that polls can be costly and time consuming while noting that the reliability of surveys can also be questioned because respondents didn’t answer for the polls or gave wrong answers.

My personal thoughts are that longitudinal surveys provide much more reliable information to those reading it. Cross sectional surveys provide information pertaining only to certain demographics of the population, not the entire population. I feel that longitudinal surveys have more validity with me because they help to demonstrate how individuals think and feel over a period of time instead of just in one moment or the next. Especially on key issues, certain things can sway an individual’s perspective in a short period of time and thus would be better represented in longitudinal surveys.

I found a good example of a longitudinal survey on the United States Department of Labor website. The survey consisted of a nationally representative sample of approximately 9,000 youths who were 12 to 16 years old as of December 31, 1996. During the screening process, an extensive two-part questionnaire was administered that listed and gathered demographic information on members of the youth’s household and on his or her immediate family members living elsewhere. The NLSY97 documented the transition from school to work and into adulthood. It collects extensive information about youths’ labor market behavior and educational experiences over time. I feel that this type of longitudinal survey helps to show information trends that exist over a period of time instead of at one given moment in time. I think this survey helps demonstrate why longitudinal surveys are more reliable and provide better information for readers to gain knowledge from.

Read Full Post »

Coleen Dickman
019:169 Response
19 November 2009
Cross-Sectional vs. Longitudinal Surveys
In his article, “Cross-Sectional Versus Longitudinal Survey Research: Concepts, Findings, and Guidelines,” Rindfleisch identifies two types of surveys: cross-sectional and longitudinal. A cross-sectional survey is conducted at a distinct point in time and measures all respondents at that point in time. Therefore, when conducting a cross-sectional survey it’s important to obtain a great degree in variance of the people surveyed. Longitudinal surveys are conducted over periods of time by following respondents as they age. Rindfleisch identifies two main issues that arise when conducting these surveys: common method variance (CMV) and casual inference (CI). Common method variance is caused by systematic method error that arises from using a single source. Causal inference, in comparison, occurs when interpreting the data and arises if a scientist is unable to find causation from the results. Common method variance blinds the researcher from making accurate inferences, and combined increase the overall error. CMV is most often an issue with cross-sectional surveys, and although longitudinal surveys reduce this bias, their practicality is often an issue. Longitudinal studies require a great degree of effort on the part of the researcher and the respondent. They are often very expensive and time consuming. Therefore, to reduce CMV bias, Rindfleisch suggests using multiple respondents in a cross-sectional survey to decrease bias by providing for more variety in responses. He also found that obtaining multiple data sources and obtaining such data over multiple periods of time would reduce CMV bias. However, if these suggestions could not be implemented in the survey, the researcher should conduct a longitudinal survey to ensure validity in their results. The APR guide further discussed these issues as well as delving into marketing research/public relations concerns.
Personally, I think that longitudinal studies are much more appealing than practical. As a researcher, having to track down the original respondents to “follow-up” presents a great number of issues. For instance; if an original respondent moves, goes on vacation, or passes away, the data is greatly impacted. I think that longitudinal surveys are extremely interesting, especially in measuring trends and how issues change in their importance with age. Yet, I still think that when looking at the bigger picture, cross-sectional surveys are most convenient.
In an article published in ScienceDaily titled, “Fifty Years and Counting: The Wisconsin Longitudinal Study,” the longest longitudinal study ever conducted is analyzed. This study conducted in 1957 in Wisconsin is the longest survey conducted and has been published over 400 times in different publications. The study called the Wisconsin Longitudinal Study, or WLS studies the entire life course of its respondents including their education, career, family, aging, and retirement. The survey has been running for more than fifty years now and will continue to do so.
Link: http://www.sciencedaily.com/releases/2007/07/070719093336.htm

Read Full Post »

Kif Richmann

Public Health Information: How Important is it?

Giving the public enough information so that they can make well informed, meaningful decisions is an important thing aspect of American democracy.  In fact, the issue of an informed citizenry dates as far back as democracy itself, having been a central issue in the rise of Greek democratic society.  And while the ability of the public to access current, accurate, and meaningful information (as well as their desire or non-desire to look for it) on political issues is important, what about the public’s ability and desire to seek information about other important aspects of everyday life?  What about, say, medical information?

In a health information survey, David E. Nelson and other scholars sought to determine just how much people know about specific health topics.  Nelson et al worked with the National Cancer Institute (NCI) to develop a major national survey that would tell them how much people know, how accurate their information is, and where they get their information from.  This survey was a spinoff from previous polls by the National Cancer Institute, and eventually came to be known as HINTS (Health Information National Trends Survey) After pretests, administration of the survey, and thorough analysis of the information, researchers determined that they had created an important tool for monitoring the public’s amount of knowledge for public health issues, as well as the quality of this information.  “Given the growing prominence and importance of health communication, coupled with changing profiles of channel usage over time, it is likely that HINTS data also will be useful for many other health application areas besides cancer.” (Nelson et al, p. 457)

Another article on a similar subject is from scholars at Harvard who conducted a survey about the public’s response to a bioterrorist attack.  This article focused on the best way to get the citizens the best information, and how this information would be used by the public.  Would they find it? Would they trust it? This study concluded that “to be effective in their communications, health officials need to know as the crisis is unfolding what Americans believe, what they know and understand, whom they trust, and what actions they are taking in response to the crisis.”

In my own experience, I think that public information is an important aspect of maintaining a healthy society.  The recent H1N1, or Swine Flu, scare has shown that the public is capable of seeking out and using the necessary information to keep themselves and their loved ones safe.  Last summer, when the virus first appeared, people were terrified of the potential epidemic. However, through the use of media and information outlets, Americans have generally been able to keep the flu to a minor incident.  A few Americans have died from the flu, but the vast majority of the H1N1 deaths have been related to previous underlining medical issues. A lot of the reasons America handled the near epidemic so well is because people quickly received information about the subject, and were able to respond accordingly. Hand sanitizers can be seen almost everywhere you go, and thanks can be given to the efficient information system for making this possible. 

But the system doesn’t always work perfectly.  As vaccinations become one of the most important tools against H1N1, not everyone agrees that it is a necessary element.  According to an ABC survey, 66% of adults say they do not plan to get vaccinated. This number is actually up 4 points from October.  And parents aren’t all on board for the vaccine either: nearly a third say they will not get their kids vaccine, according to the poll.  But it seems that most people do trust the information, as the ABC survey found that over 60% reject the idea that the government exaggerated the danger of H1N1.  But what about the almost 40% who said otherwise?  In general I think this polls is a good example of the general population, and also shows how important good information is when it comes to health issues.

Read Full Post »

Keely Jarvill
Quinjiang Yao
Introduction to Public Opinion
18 Nov 2009
Cross Sectional Surveys vs. Longitudinal Surveys
When having a profession in the public relations field many times they have to deal with polls and trying to analyze the results for their clients. Many public relations professionals use the surveys to predict marketplace behaviors. Given that the PR professionals spend so much time looking at surveys and polls, improving the validity of them would be a wise decision. In the selection “Cross Sectional Versus Longitudinal Survey Research: Concepts, Findings, and Guidelines” the authors explain that improving cross sectional surveys would be a good start to help the PR agents and offer longitudinal surveys as a replacement.
The authors explain in this selection that many people are worried about cross sectional surveys validity because of common method variance and low casual interference, which can lead to bias. The authors explain to help reduce this problem is to employ multiple respondents, obtain multiple types of data, and gathering data over a longer period of time, which can be defined as a longitudinal survey. The difference between cross sectional surveys and longitudinal surveys is that cross sectional takes place at a single point in time and looks at one population, whereas longitudinal surveys take place at multiple periods of time and looks at many different populations. Even though longitudinal surveys may be more effective and reliable, they also can tend to be more costly and time consuming. That is why many companies stick with cross sectional because they are easier to gain information and less costly. They explain the Monte Carlo approach might help the bias in cross sectional surveys if the magnitude is at a high level. If the predictors and outcomes are at a low in cross sectional surveys longitudinal surveys can help that problem out by enhancing the CI. In the APR selection it discusses the drawbacks and advantages of polls, which can make participation easier and people’s participation can remain unknown. It also discusses how it can be costly, time consuming, and the reliability of the responses can be biased because the respondents didn’t answer their phones or gave a wrong answer.
What my personal stance is on surveys is that I would prefer to read information out of a longitudinal survey. I feel that the survey provides more reliable information. Cross sectional surveys only provide information about what one certain type of population thinks which does not represent all the population. Longitudinal surveys can give me information on what many populations think over a long period of time. Considering many people change their opinion in a matter of days, I feel that longitudinal surveys would provide better results.
I found a good example of a cross sectional survey on bmj.com. This poll conducted by the company only interviewed 14-17 year olds on the issue of the awareness of the tobacco marketing techniques. The results were that 65% of 15 year old recognized tobacco ads and 35% of 16 year olds recognized ads. This is a good example of how a cross sectional survey can tend to be biased because most of the respondents surveyed do or had tried smoking. If you are a smoker more than likely you will look for ads to try to save you a deal or a new kind of product coming out. If this type of survey was longitudinal they could have less smokers because of the more population interviewed.
Cross sectional surveys and longitudinal surveys have many different comparisons. Longitudinal surveys interview many different people over a longer period of time and cross sectional surveys interview the same population over a shorter amount of time. Both surveys have their disadvantages, but their main goal is to help get the public’s opinion projected in society.

Read Full Post »

Darin Evangelista
Reaction Paper
11/19/09
Advantages and Disadvantages of the Survey
The reading for today focused on the validity of two different types of surveys, both cross-sectional and longitudinal surveys. The researchers note that the problem with cross-sectional surveys is they do not do well to limit a common method variance bias and they do little to prove causality. The reason for this is that a single respondent at a single point most often completes the survey in time, making the study “prone to potential CMV bias.” Longitudinal surveys are often used to combat these problems because “temporal separation reduces the cognitive accessibility of responses to predictors collected at an earlier time, which in turn reduces the likelihood that these earlier responses will influence subsequent responses to outcome variables.” Longitudinal surveys are not always fully relevant due to the time the survey is taken. Changing time and different events can change respondent’s opinions.
The APR study guide outlines the necessary steps needed to conduct a successful survey. In order to properly conduct a research survey the researches must follow four basic steps, first they must research, then plan how the survey will be conducted, then it is time for the actual implementation of decided survey methods, and lastly is the evaluation of the data collected. In this final step the researchers will develop any conclusions that can be drawn from the data, like why this information is important or how it changes ones thinking on a certain topic. The study guide also explains various types of survey methods and tells of the positives and negatives of the survey. The guide then goes on to explains different methods and terminology used by surveyors, like the use of focus groups in information on sampling.
In high school I was required to take a survey every year about various lifestyle choices. Questions included things like my aspirations to drug use to family life. I think the reason these surveys were conducted in my school was so researched could try to correlate peoples actions with another aspect of their life. For instance, if ones parents were divorced, maybe they are more likely to smoke cigarettes or do badly in school. The problem I had with these surveys is that they were often times too long. By the third or fourth page I began to lose interest in the questions and started to answer randomly. I know many of my peers did the same thing that would greatly reduce the validity of the surveys.
An ABC News Poll that I looked at studied the American publics opinion on the state of the economy over the past year. The study showed how the general thought on the economy is higher than it was in November 2008, but it was not a steady increase. Instead the longitudinal polling allowed readers to see the ups and downs of the publics’ opinion. This method is interesting because it lets the reader note times when peoples thoughts on the economy were at highs and lows, then the reader can look back at events that occurred during those times and see how they effected public’s general outlook.

Read Full Post »

Kahle Novak

November 19, 2009

Reader Response #6

 

                                                Polls and Marketing Research

 

            When it comes to surveying managers in a business, marketing representatives want to understand, explain, and predict the kinds of behaviors that will exist in the marketplace now and in the future based on a questionnaire.  Although, recently it has become an issue of how valid the results of a survey research really are.  “Two issues dominate these concerns (1) common method variance (CMV) (i.e., systematic method of error due to the use of a single rater or single source) and (2) casual interference (CI) (i.e., the ability to infer causation from observed empirical relations) (Rindfleisch, p. 261).”  This is the main concern in the findings of the marketing research taking place of a cross-sectional survey versus a longitudinal survey.

            Cross-sectional is the current survey method but in order to reduce the common method variance and the casual interference, collecting longitudinal data would be beneficial.  Cross-sectional surveys are “completed by a single respondent at a single point in time (Rindfleisch, p. 262).”  Longitudinal is where data can be collected over multiple periods of time.  In order for a reliable survey, researchers have recommended collection strategies including: “(1) employing multiple respondents, (2) obtaining multiple types of data, or (3) gathering data over multiple periods (Rindfleisch, 262).”  It is tested and believed that longitudinal surveys are a better method in collecting information.  Although, in the results we find that cross-sectional surveys and longitudinal surveys are both adequate means of measuring information.

            Recently, when collecting data for our polling experiment, we started the research by handing out our survey questions to a single person and allowing them a single point in time to fill out their survey.  This did not include more than one person filling out the same exact survey, which would include two different opinions, and it was given to them at one point in time, rather than asking them the same questions at a later point in time.

            One specific example of a cross-sectional survey was investigated by Robert Quinn and Graham Staines in their “Quality Of Employment Survey, 1977: Cross-Section.”  In this study, 1,515 workers, ages 16 years and older who worked a paid 20 hours or more a week job in the United States in 1977.  This survey was used for purposes of understanding the working conditions in the American labor force.  The quality of employment, the problems within the workforce, and job satisfaction were some questions in the survey just to name a few.

            An example of a longitudinal survey is investigated by the United States Department of Commerce and the Bureau of the Census of the “Survey of Income and Program Participation (SIPP) [1984 Panel].”  This survey was a way to examine the economic well being of people and households in the United States.  The three basic elements in the survey include a control card, which is used to record the social characteristics of each person, a core portion of the questionnaire including financial questions, and supplemental questions during visits to people’s houses.  The difference among these two studies includes the type of study and how they go about taking the survey.  The first on has a one time questionnaire while the second one has thee different elements that occur at different times and look at other variables.

 

“Quality of Employment Survey, 1977: Cross Section.” Description and Citation—Study No.

7689. 1984. Inter-University Consortium for Political and Social Research. 18 November 2009. <http://icpsr.umich.edu/icpsrweb/ICPSR/studies/7689/detail?archive=ICPSR&amp;

q=cross+sectional+survey>

 

“Survey of Income and Program Participation (SIPP) [1984 Panel].” Description and

Citation—Study No. 8317. 1985. Inter-University Consortium for Political and Social Research. 18 November 2009. <http://icpsr.umich.edu/icpsrweb/ICPSR/studies/8317

/detail?archive=ICPSR&q=public+relations+survey&keyword=economic+conditions&paging

.startRow=1>

Read Full Post »

Older Posts »