Surveys

From EvaluationWiki

Jump to: navigation, search

Contents

Types of Surveys[1]

Surveys can be divided into two general categories on the basis of their extensiveness. A complete survey is called a “census.” It involves contacting the entire group you are interested in -- the total population or “universe.” The other category is more common; it is a sample survey. A sample is a representative part of a whole group (universe). Thus a sample survey involves examining only a portion of the total group in which you are interested, and from it, inferring information about the group as a whole.


Advantages and Disadvantages of the Two Types of Surveys

One of the decisions to be made in surveying is whether or not to sample. Parten[2] presents a list of advantages and disadvantages of the sample survey. (These, in turn, imply the advantages and disadvantages of the census.) The three most important considerations for the surveyor are: speed, low cost, and increased accuracy and analysis of the data.


By sampling only a small portion of a large population, it is possible to collect data in far less time than would be required to survey the entire group. Not only is the data collection quicker, but the data processing and analysis also require less time because fewer pieces of data need to be handled. Rapidly changing conditions and the short turn-around time imposed in many surveys make the efficient use of time a critical variable. If an accurate snapshot of the attitudes of a particular group is desired, currency is of paramount importance. Professional political pollsters make their living by providing quick snapshots of the “political climate.” Results of such polls lose their accuracy very quickly (sometimes in as little as 24 hours--particularly in the days preceding a major election). So, for these pollsters, time is truly of the essence. It's probably a safe bet that those of you reading this guide will not need that degree of speed. Nevertheless, speed is essential to ensure the data are "fresh," especially when it comes to assessing public opinion in a volatile or contentious area before they change appreciably.


The smaller amount of data gathered by sampling as opposed to surveying an entire population can mean large cost savings. By limiting the group to be surveyed, less time, hence less cost, are involved in collecting, formatting, and analyzing the data. In addition, if automated data processing (ADP) equipment is being used to analyze data, your overall time investment will be even less, as will be the overall cost. Sampling allows you to do a credible job for a smaller investment of time and money. Parten[2] also notes that sampling enables the surveyor “to give more attention to each return received and to make certain that the data are as accurate as possible”. This attention may lead to more precise information than would a less careful collection of data from the entire population. Nothing more than a rudimentary quality control is possible for the great volume of raw data gathered in a census. The more data collected, the greater the potential for making “accounting” errors. The disadvantages of sampling are few, but important. The main disadvantages stem from risk, lack of representativeness, and insufficient sample size, each of which can cause errors. Inattention to any of these potential flaws will invalidate survey results.


It is important to realize that using a sample from a population to infer something about the entire population involves a risk. The risk results from dealing with partial information. If risk is not acceptable in seeking the solution to a problem or the answer to a question, then a complete survey or census, rather than a sample survey, must be conducted. Determining the representativeness of the sample is the surveyor's greatest problem when sampling. By definition, "sample" means a representative part of an entire group. To avoid the charge of using “biased data,” it is necessary to obtain a sample that meets the requirement of representativeness, and this is not an easy task. Without a representative sample, a survey will, at best, produce results that are misleading and potentially dangerous.


The final major problem in sampling is to determine the size of the sample. The size of the sample you need for a valid survey depends on many variables including the risk you are willing to accept and the characteristics of the population itself. Here, it is sufficient to say that if sampling becomes too complicated, or the required sample size becomes too large, the easiest solution may be to survey the entire population. The decision as to whether to survey the entire population or only a sample of it is not based on the above advantages and disadvantages alone. It is affected by many other variables.

To Survey or Not to Survey

Before attempting a survey, you should investigate some basic facts and answer some pertinent questions. The result of this investigation will be a greater realization of the work involved in producing a survey. Perhaps it will lead to a decision not to survey.


Surveys demand time (maybe more time than you have available). The exact amount of time varies greatly from survey to survey depending on the number of people to be surveyed and the content of the survey. A survey of a few questions administered to the people in your office may take only a day or so, whereas a larger survey administered to a great number of people located worldwide can take over three months from the time the survey is delivered to the printer. And this does not include the time needed to design the survey and construct the questionnaire. Moreover, coordination with officials and the customers of the survey takes additional time. If your estimate of the time needed to produce the survey exceeds your deadline date, you are likely to decide you do not have the time to conduct a survey. A hurried survey wastes both your time and that of your respondents. The results of a hurried effort are questionable at best.


Surveys are expensive to produce. The solution to the problem or the answer to the question may not be worth the cost to produce it. Even if it would be worth the price, you may not be able to obtain the needed funds, either from your own pocket or from your organizational budget. Although no standard estimates of survey cost are available, some of the items of expense can be examined. The primary expense is in time and effort; the time you spend producing the survey could be spent on other tasks. If other personnel are needed, they will have to be paid. Access to typewriters, word processors, and calculating machines (or computer resources) is a must. If you expect to gather a great deal of data, the cost of renting ADP time and of purchasing ADP scanner sheets should be examined. Surveys of more than 150-200 respondents cannot feasibly be tabulated by hand. The same is true for groups of less than 150 respondents if the survey questionnaire is lengthy. The final cost involves supplies. At a minimum you will need paper and envelopes. You may also have to pay either the cost of printing the survey questionnaire or the postage or both. Each of the above costs that applies to your survey should be estimated and the total cost measured against the survey requirement.


Since surveys are being used more and more, the information you want may have already been gathered. So before you undertake a survey, first make sure the answer to your problem does not already exist. Next, evaluate the time you will need and determine the cost involved to produce the survey results, and then weigh these findings against the importance of the survey. Undertake a survey only if it is worth the time, effort, and cost to make it a good one.


Developing the Purpose, Hypotheses, & Survey Plan

The first steps in producing a survey are the most important. They determine where you are going (the purpose), how you will know when your are there - or what you expect to find (the hypotheses, objectives, or research questions), and by what route you will go (the survey plan). If these steps are not well planned, all the remaining steps will be wasted effort.


The Purpose

The first step in producing a survey is to define the purpose or objective of the survey. “A clear statement of purpose is necessary not only as a justification/explanation of the project, but also as a guideline to determine if future actions in the project are in support of the original purpose”.[3] Without knowledge of the exact nature of the problem (objective), you cannot decide exactly what kind of data to collect or what to do with it once you have it. Usually a staff officer is given a problem or objective; it seldom originates with him. But this does not relieve the individual of responsibility for insuring that:

  • the problem is well stated
  • the surveyor understands exactly what the problem is
  • the stated problem is the real problem


The survey should be designed to answer only the stated problem. Adding additional interesting objectives will lengthen and complicate the survey while clouding the real issue.

The Hypothesis, Objective, or Research Question

Once the problem has been clearly stated, the next step is to form one or more hypotheses. The hypothesis is actually your educated guess about the answer to the problem. It should not be a capricious guess, however. It ought to be based on your prior experience related to the problem, or perhaps any knowledge you may have of previous research done on the topic. Without such a framework in which to make an educated guess, you really have no basis for making a guess at all. If you do not have a clear basis for formulating a hypothesis, you should instead develop one or more objectives or questions to frame the scope of your questionnaire.


For example, if a problem is identified on the base as declining use of the Officers' Club, an immediately obvious question comes to mind: “Are the officers on this base satisfied with the Officers' Club facilities?” This would be suitable as a research question. It is possible, though doubtful, if you could come up with a supportable hypothesis, or educated guess, as to the answer to the problem. You may, for instance, have gathered some anecdotal evidence (overhearing colleagues talking) of dissatisfaction with the club's facilities. But, this may not be sufficient for making an educated guess that this is the real reason for the decline in club use. The problem could be seasonal; it could be related to a decline in the officer population on the base; or a number of other possibilities. The point is that without some credible evidence to support a hypothesis, you should probably not formulate one.


If you formulated a hypothesis for the current example on the basis of the anecdotal evidence available to you, you would naturally construct a questionnaire to survey the opinions of officers regarding their use, or lack thereof, of the Officers' Club and the reasons for it. You might never think to gather data from the base military personnel office to see if the officer population is lower now than usual or if there are seasonal (cyclic) trends in the size of the officer population on the base. In other words, establishing the hypothesis may blind you to collecting data on other possible causes of the problem. This is why all researchers are cautioned not to formulate hypotheses unless they have a solid base in theory or previously gathered evidence that suggests the hypothesis is, in fact, probable.


Hypotheses must be carefully written. They should not contain moral judgments or biased statements such as “All pilots are good leaders.” There are many ideas on what constitutes a good leader and your idea may not be the same as those of the people you will contact. Avoid words like should, best, good, bad, and ought.


Hypotheses should be as specific as possible. Avoid words such as most and some. If by most you mean a majority, then say majority. A survey can more easily be designed to test whether “more than 75 percent approve” than whether “most approve.”


A well-formulated hypothesis, objective, or research question translates the purpose into a statement that can be investigated scientifically. The level of difficulty you will face in producing a valid survey will increase dramatically if they are not well formulated. Take care in doing this step, and it will save you much effort later in the survey development process.

The Survey Plan

The next step after determining the purpose and hypotheses is constructing the survey plan. The purpose of the survey plan is to ensure that the survey results will provide sufficient data to provide an answer (solution) to the problem you are investigating. The survey plan is comprised of three different parts:


  • data collection plan
  • data reduction and reformatting plan
  • analysis plan


None of these plans stands on its own. Decisions you make on how you will analyze your data will affect your data collection plan. The type of data reduction you do will affect not only the types of analyses you can do, but also the amount and types of data you need to collect. Because these plans are closely interrelated, they should be developed concurrently.

The Data Collection Plan

The purpose of the data collection plan is to ensure that proper data are collected in the right amounts. The appropriateness of the data is determined by your hypothesis and your data analysis plan. For example, if you plan to analyze your results by age group to test a hypothesis, then you must collect data from each age group whose opinions you want to know. The right amount applies to sample data. The use of sample data involves risk, and the amount of that risk is determined by the size of your sample. The amount of risk you are willing or able to accept should be stated in your analysis plan. Proper and right come together when your analysis plan involves both sampling and analyzing data by groups. You not only have to collect data from some members of each group you plan to analyze, but you also have to see that each group provides a response rate that is high enough to ensure your meeting your minimum risk level.

The Data Reduction and Reformatting Plan

The purpose of the data reduction and reformatting plan is to identify up front and to decrease as much as possible the amount of data handling (reduction and reformatting) you will have to do. This plan is highly dependent on the other two plans. As previously mentioned, if your collection plan calls for a great deal of data, you should plan to use a computer to analyze the data. If ADP scanner sheets are to be used to record respondents' answers, include the sheets with the questionnaire so the respondent can fill out the scanner sheet. This will save a great deal of time that you would have to spend if you transferred the survey data to the scanner sheets yourself. It also eliminates the possibility of your making errors in transferring data. You should coordinate in advance with the ADP personnel to make sure they will be able to scan your answer sheets and, if necessary, analyze your data within your timeframe. ADP shops are busy places. The prudent surveyor will “book” the scanning and analysis jobs well in advance with ADP personnel to ensure their resources are available when needed.


A strong potential for error and tedious corrective work lies in data reduction and reformatting. Proper care in developing this plan can save a great deal of time later and preclude error.

Open- and Closed-Ended Questions

The use of Automatic Data Processing (ADP) necessitates the use of closed-end questions -- a type of question you should consider even if you are hand-tabulating your data. A closed-end question lists possible answers from which the respondent picks the one he/she likes best. An example is the common multiple-choice question. The open-end question is one to which the respondents write the answer out in their own words. At first glance, the open-end question seems superior since respondents supply their answers rather than ones from your list of answers. But the wide variety of answers respondents generally provide to open-end questions turns out to be a great handicap later. For every open-end question, there are virtually an infinite number of possible answers. Since you cannot analyze an infinite number of answers, you must devise some means of categorizing this diversity of answers into a smaller, more manageable group. You will find yourself spending a tremendous amount of time reading, comparing, categorizing, and recording each answer. Much of this time can be saved if you use care in developing the questionnaire and constructing your own categories in advance. Construct each question so that every possible major category of response is contained in the answer list.


Then, later, all the computer will have to do is count the number of answers in each category. By having the survey respondent, not you, categorize the answer, you will collect data that is more valid, reliably, and accurate than if you did the categorizing yourself.

Analysis Plan

Finally, an analysis plan ensures that the information produced by the analysis will adequately address the originally stated hypotheses, objectives, or questions. It also ensures an analysis that is compatible with the data collected during the survey. In the analysis plan, you determine which statistics you will use and how much risk you can take in stating your conclusions. Each of these decisions will affect the amount and type of data you collect and how you will reduce it. Novice researchers often misuse statistical analyses out of ignorance of the assumptions on which the statistics are based. The most often committed error in statistical analysis by novices is using a statistical technique with inappropriate data. The results of such analyses appear to be legitimate, but are actually impossible to interpret correctly.

Concluding Thoughts

Oppenheim[4] suggests that to make sure all these parts of the survey plan are correctly interlocked, you can simply approach the natural sequence of survey operations in reverse order. First determine what conclusions you are interested in; then decide what statistics and results will be needed to draw these conclusions. From this, the type of questions needed and the nature of the sample can be determined.


A conscientious survey plan will help you produce a well designed survey. The proper data will be processed correctly and efficiently to produce the information required to shed light on, and hopefully provide a solution to, the original problem.

References

  1. Ross, Keneth C. (1996). Air University Sampling and Surveying Handbook: Guidelines for planning, conducting, and organizing surveys.Retrieved September 9, 2006, from http://www.au.af.mil/au/awc/awcgate/edref/smpl-srv.pdf.
  2. 2.0 2.1 Parten, M. (1950). Surveys, Polls, and Samples: Practical Procedures. New York: Harper and Brothers, p 109-110.
  3. Department of the Air Force. (October 1974). A Guide for the Development of the Attitude Opinion Survey. Washington, D.C., p 2.
  4. Oppenheim, A. N. (1966). Questionnaire Design and Attitude Measurement. New York: Basic Books, Inc.
Personal tools