Survey Data example essay topic

... ay. The respondents may usually be located at home during the evenings while during the day; they are at their places of work. The accuracy of a survey can be improved by drawing larger sample, while obtaining less than the required minimum number of samples may severely impair the results. In the field, it is difficult to retrace one's steps to re sample individuals from areas that have already been covered, and frequently individuals pre-selected may not be available. Therefore, it is a wise practice to over-sample in advance, to utilize the survey team's time to the maximum extent possible. Also editing the information sheets during slack time in the day is time saving and very worthwhile.

It is much easier to backtrack locally for additional data. Incomplete questionnaires may have to be discarded. Scheduling A detailed work schedule for completing each major step of the survey and its analysis must be prepared at the outset in order to complete the survey in time. CONDUCTING THE INTERVIEW Before starting the interview, make sure that the interviewee is the right respondent. Verify responses for accuracy by cross checking. Often, individuals misunderstand what the researcher is asking, or only tell what they think the researcher wants to hear.

They may be trying to impress the researcher, gain his sympathy, or avoid discussing the topic at all for lack of knowledge or fear of embarrassment. Do not promise anything, except to pass on the information, unless there is an authority to take corrective action. The individual being interviewed may regard the researcher as a representative of the government who can and should do something about a local problem situation. Idle promises will only result in a lack of confidence and reduce cooperation the next time around.

Use judgment whether to follow a structured questionnaire format reading off each item; or an unstructured interview style using the questionnaire as a checklist, but employing a lot of additional extemporaneous probing questions. The structured style may get a response to every answer, but may scare or inhibit the response, especially if the answer is recorded in the presence of the person being interviewed. On the other hand, some people feel more important when they see that what they say is being written down, and often think that if their responses are not written down, the interviewer may forget and / or fail to pass on their comments. Unstructured interviewing generally leads to a wide-ranging discussion.

It may take longer but may gather more supplementary data that may also be useful. However, it is not generally possible to statistically analyze such additional data. Sometimes, it is critical that every respondent be given only the precisely formatted question, so that responses are standardized. Extemporaneous questioning often introduces interviewer bias. Shao (1967) provides the following strategies which should be observed when conducting an interview: 1. Introduce one-self.

2. Make sure that the interviewee is the right respondent. 3. Put the individual being interviewed at ease. 4.

Tell the reason for the survey and its use. 5. Tell the individual how he / she was selected to be interviewed. 6. Assure him / her of confidentiality or anonymity of results. 7.

Tell him / her how long the interview is likely to take. 8. Ask if the time is convenient for an interview now. 9. See whether there is a suitable place to conduct the interview.

Privacy is often desirable, especially when asking personal questions. ANALYSIS OF DATA After the data have been gathered and recorded, they must be edited, processed, and interpreted. Editing Prior to use, raw data gathered must be screened using consistent guidelines. The principal purposes of this are to review for clarity, internal consistency, correction and mark-up for further processing. Preliminary editing in the field should speed up this process. Clarity.

Data recorded under field conditions are sometimes almost illegible and / or unintelligible. Numbers may be illegible, and many cryptic comments may have been added to the standardized responses which might qualify the answers recorded from 'Yes' to 'Yes, but... ' (Smith, 1981). Whenever possible, items that are questionable should be reviewed with the individual making the survey. However, the one conducting the survey cannot always read his / her own writing, and / or does not recall the context in which the comments were made, even though they may have seemed meaningful at the time.

Where multiple choice responses have not been used, developing a standardized scheme to classify open-ended comments received is an extremely difficult task. Sometimes, there are things that were overlooked, or thought not to be important when designing the questionnaire, actually have great significance while other questions may no longer be pertinent. Thus, some preliminary modification (or even elimination) of questions and responses may be necessary. This emphasizes the need to carefully plan and structure the survey before gathering the data, not afterwards.

Internal Consistency. Check marks may have been placed in more than one option of multiple choice questions even though it was originally specified that only 'one of the above' was to be checked. There may be clarifying comments in the white space as to why, or there may be no explanation at all. With number responses, editing is frequently required to recalculate the recorded values into the standardized units requested.

Correction. Decisions have to be made on how to treat questionable data. Should the data be rejected outright as erroneous; counted at face value regardless of its apparent error; or retained but reduced in value, with an attempt to figure the intent? This is all part of the editorial task. Mark-up.

Finally, to simplify the data processing task which follows, it may be necessary to transform the check marks in the standardized responses into base numbers. Calculating the Data After editing preliminary calculations and double-checking, the questionnaires can be ranked in descending (or ascending) order for each question, and then the data can be transcribed. From this, a frequency distribution can be constructed and the mean, standard deviation, standard error, and confidence computed. When the information has been computed for one question, the entire process should be repeated for the next questions until all the questions have been calculated. The entire database should be reviewed, rechecked, and recalculated after a suitable time interval to assure that it is as correct as possible. Then, the data should be analyzed and interpreted for comparisons, trends and significance with the use of tables, graphs, etc.

Findings should be reviewed before proceeding. PRESENTATION OF RESULTS The final step in the survey process is to present the findings of the study. This is a very critical phase. In fact, it is the point of the whole exercise.

Designing questionnaires, interviewing, and statistical manipulations of various kinds were just means to the end. Many well conceived, planned, and executed surveys fail miserably because they do not communicate with their intended audience. Remember that readers have not had the experiences in traveling, interviewing, researching and analyzing survey data -- so it is difficult for them to empathize with the researcher. They will only know what researcher tells them. It is imperative that they get the message loud and clear. The first principle of report writing, therefore, is to purge drastically.

The second principle is to simplify what is left, and then summarize. If details must be included because they are too precious to throw away, consider putting them in an appendix in which other researchers may delight to wallow. Above all, provide the readers with a page or two of summary of the purpose of the study, the findings, and conclusions. Presentation is a whole subject itself. This paper is, therefore, limited to the following major points in writing survey reports: Avoid technical jargon, abbreviations and acronyms unless the intended reader is completely familiar with them.

Round off numbers wherever possible, it does not usually distort a thing. Where tables are used, get all the data on one page whenever possible. There is nothing that will distract a reader from gleaning the message from the table, more than having to flip pages. Tables should be organized so that a single message is highlighted. Comprehensive matrices of basic data are only useful for researchers to analyze -- they do not communicate until they are interpreted.

If there is a need to include a comprehensive table, the appendix is the place for it. Extract from it the point to be made, and then prepare a condensed version in the text at the appropriate point. After using a table, summarize in the narrative what the reader is supposed to learn from studying it. Some people have a mental block against numbers and only read the text skipping over tables. If there is a need to go into the detail on a point, and it would clutter up, then use a footnote. Remember, however, that a footnote is best seen at the foot of the page on which the point is raised.

Footnotes, relegated to the back of the text, rarely get read in relation to the points they are clarifying. Obviously, there is much more to the subject than is contained herein. A number of topics worthy of extensive treatment have been simplified and summarized, while others have been completely ignored. In doing this, the writer tried to follow the 'mini-skirt' principle -- keeping it long enough to cover the subject, and at the same time, short enough to remain interesting! LITERATURE CITED Anderson, R.L. and T.A. Bancroft. Statistical Theory in Research.

McGraw-Hill Book Co., Inc., 1952. Shao, Stephen P., Statistics for Business and Economics, Ohio, USA: Charles E. Merril Publishing Co., 1967. Smith, Kenneth F., Applied Survey Methods for Development Projects, The US Agency for International Development Training Center, 1981. - -.