It is very important not to ask too many questions! This bores the respondent and increases the risk that the respondent will skip the last questions. Maybe those questions were essential? A basic requirement for the questionnaire is to ensure a clear communication, i. Keep it simple! Keep in mind that the respondent is not an expert! If the results of the questionnaire are to be compared with results from other questionnaires, care must be taken to ensure comparability. This means that you pose the same questions and use the same answer categories as in the other surveys.
Remember to always ask about one thing at a time. It may be tempting to put two issues together in one question, but the result is rarely good. An example from real life is the following question: Are you shopping — Alone? Here, the researcher is dealing with two issues: a Whether the respondent is responsible for a large or small proportion of purchases in the household. The purpose of the question was to identify the person responsible for most of the shopping in the household.
It is always a good idea to test the questions on a few people in order to see if they are understood and are answered without problems. Such a study known as a pilot study can detect many problems before it is too late! There may be errors that come from the question wording and errors arising from the data collection.
This can be — Unclear questions. This means that some respondents do not participate in the survey. This may be due to problems in the data collection. Non-response may mean that the results of the survey are misleading! The two most frequent causes of non-response are: — People are not at home, or not reached by telephone interview or visits. The only solution to not reaching people is to make numerous attempts of contact telephone calls, visits, etc.
Typically you should make attempts both in the weekend and on weekdays, both in daytime and evening hours, if you want to increase the response rate. This means at least four contact attempts, which of course is costly The problem with refusal is growing as the number of more or less dubious marketing companies increases, which again increases the number of phone calls to each household.
Clinical Statistics For Non-statisticians
People have simply had enough! The problem is of course impossible to avoid completely. This again increases the costs, however. Usually, there is a relationship between cost and quality! The above list is roughly arranged in order of increasing cost and quality. Thus Internet or e-mail questionnaires are cheap, while personal interviews are expensive.
The quality depends largely on the non-response rate, i. The non-response rate is considerable by Internet or email; by personal interviews it is much smaller. Errors in data are also a quality issue. The traditional mail questionnaire gives no opportunity to correct errors during the interviewing process, as do the other data collection types.
STATISTICS FOR ALL
The first two methods of data collection are carried out without an interviewer, while there is an interviewer present with the last two methods. The interaction between interviewer and respondent in the last two data collection methods has both advantages and disadvantages that can affect quality: — Advantage: The interviewer can help the respondent, if there is any doubt about what is meant by a question. This risk is largest in personal interviews, but can be minimized by careful training of interviewers.
The last two data collection methods involve a direct contact between interviewer and respondent. The talented interviewer can exploit this, if a respondent is becoming impatient. In personal interviews different visual effects, which can help the respondent to understand a difficult issue, can be shown.
- Statistics for Non-Statisticians Webinar - AATB!
- Dragon Cafe.
- Current students.
By telephone interview audio files can 12 1 Data Collection be played during the interview. In mail questionnaires one can display pictures. Often you will use a combination of two data collection methods. One collection method often telephone is used for most interviews. For the reminders is used a second data collection method, e. The skilled interviewer can better persuade a respondent to participate in the survey.
This can be used to achieve a higher response rate. Also, this enables an assessment of whether there is a difference in the results among those who answered respectively did not answer in the first round. This provides information on the validity of the survey results. This register contains contact information name, address, etc. We will discuss the process of selecting a sample in Chap. Fitness Club chooses to send a mail questionnaire to all the selected customers. It is a relatively inexpensive solution. A disadvantage is that for instance the questions about their health are answered subjectively.
Shamshuritawati - Statisticians
This is, indeed, a common feature of most questionnaire surveys! Another drawback is that such surveys often lead to high non-response rates. To reduce this problem, Fitness Club encloses a stamped addressed envelope. Also, the customers responding will participate in a lottery, where they can win smart mobile phones and other electronic gadgets, which are particularly attractive to kids of this age group. Chapter 2 Presentation of Data In this chapter we show how to present the results of a questionnaire survey, using graphs and tables. Graphs charts, plots, etc. They are also useful tools to find unlikely e.
Here, only the main types of graphs are covered. There are many other types of graphs than those shown here. Tables are another way of presenting data, which is also discussed briefly here. They conveniently summarize information from a table in a clear and illustrative manner. As one focus of the young customers is weight loss, a table of average height and weight by sex is interesting. It may, in tabular form look as shown in Table 2.
antrusnala.cf A corresponding bar chart is shown in Fig. We see, at a glance, that the boys are both slightly taller and slightly heavier than the girls.
- Starting with Hobbes (Starting with...).
- The Ghost of Black Rock Castle (The Jess Mystery Series Book 1)?
If we consider for a moment only the average height, it could in graphical form look like the one shown in Fig. This chart contains the same information on the height of girls and boys, as the graph above.
However, most people will get a wrong impression of the situation by considering this chart. The boys seem to be much taller than the girls, because the lower part of the bars is cut off! It is therefore important to be aware of the axes when studying bar charts, as well as when constructing them.
Forgot your password?
This means: in the interval from just over to cm there are in total three kids. The intervals must of course be constructed so that they do not overlap. Therefore, only one endpoint belongs to the interval.
The center of the first interval is cm. Counting of the frequencies can be done manually. When the frequency data from the table are plotted in a bar chart, it looks like this as shown in Fig.
The general considerations you should have in connection with determining the number of bars and the width of the bars are: — The graph should fit onto the paper or screen. Be aware of the minimum value and the maximum value. There must be no doubt as to which interval an observation should belong. You must be sure to which interval the endpoints belong. As a rough guide, you can use Table 2. The number of bars should be between 3 and 7, so 5 is probably a fairly good choice. Technical Note To determine the number of bars, we can use the following formula: No.