問卷法 Questionnaire ppt

1、Questionnaire,Duan Lifu 2013.12.16,Content: 1. Definition 2. A brief description 3. The main parts of a questionnaire 4. The design of questionnaire 5. Scaling techniques 6. Selecting subjects 7. Administering the questionnaire,1. Definition,Questionnaires are any written instruments that present re

2、spondents with a series of questions or statements to which they are to react either by writing or their answers or selecting from among existing answers. Brown, 2001:6,The term “questionnaire” refers to an instrument for the collection of data, usually in written form, consisting of open/closed que

3、stions and other probes requiring a response from subjects. Nunan, 2002: 231,A survey study usually involves a relatively large sample and is descriptive in nature. It either describes the features of a few variables or the relation between two or more variables. A survey study may be cross-sectiona

4、l, when data are collected from different people at the same point in time; it may be longitudinal , when data are collected from the same group of people at different points over a period of time,2. A brief description,3. The main parts of a questionnaire,Title Instruction : Instructions are of two

5、 types: - general instruction at the beginning of the questionnaire - specific instructions introducing each new task (3) Questionnaire items (4) Additional information (5) Final “thank you,4. The design of questionnaire,4.1 Criteria for a good questionnaire: (1) Having high internal validity By say

6、ing a questionnaire should have high internal validity means that items in the questionnaire must measure the variables you really want to investigate. But some researchers might fail to meet the criteria because of lack of skills or experience. Actually, having low internal validity is a common fla

7、w occurring in a questionnaire constructed by a novice researcher,A MA student wanted to investigate students degree of anxiety in speaking L2, the following three items were part of her questionnaire: I try to avoid answering questions in class. I try sit in the corner of the classroom so that I wi

8、ll not be asked questions frequently by the teacher. I do not like to speak in the public,4.2 Taking a professional outlook How a questionnaire looks can affect the attitude of the respondents which in turn influence the quality of their response. You should avoid clipping or stating papers together

9、 if a questionnaire contains more than 5 pages. One question should not be printed across two page. You should avoid crowing questions together to make the questionnaire looks shorter. The paper for printing should be of high quality and the printing itself must be clear and easy to read,4.3 The pro

10、cess of designing a questionnaire Selecting the mode of administration. Specifying what kind of data you intend to collect. Determining the way you process the questionnaire data. Deciding on the content of individual item. Choosing questions structure. Determining the order of questions. Deciding t

11、he format of the questionnaire. Conducting a pilot study to test the questionnaire,1) Selecting the mode of administration. Two common models: By person: A group of subjects usually answer the same questionnaire together. The organizer gives the respondents instructions on how to respond to the ques

12、tionnaire and clarifications if there are any ambiguous items. The organizer can also check whether the respondents has responded to all the items and thus the rate of data can be greatly reduced,By mail: If the respondents are spread out in different cities or even in different countries what you c

13、an do is to mail the questionnaire to the target people. Researchers cannot control the overall response rate of missing items. Even if a questionnaire is returned ,s/he cannot be sure that questionnaire is filled out by the people s/he expected and under which circumstances the questionnaire is com

14、pleted,2) Specifying what kind of data you intend to collect. At this stage, you have to understand what kind of data you really want to collect. There are five type of information. Experience/behaviour (what a respondent does) Opinion (what a respondent thinks) Feeling (what a respondent feels) Kno

15、wledge/abilities (what a respondent knows and what is he able to do) Background (such as date of birth, age, gender, martial status,Novice researchers cannot clearly differentiate belief items from behavior items. The structure of belief item is “I think/ believe that” I think learning an L2 well re

16、quires painstaking effort is to find out the respondents opinion or view rather than behavior. I put a lot of effort into learning an L2 is to investigate students behavior rather than the view,3) Determining the way you process the questionnaire data. Individual- item-based (單個(gè)題目問卷) If the basic un

17、it in the data analysis is an individual item, such as describing frequency and the mean of one isolate item, then such a questionnaire is called individual- item-based. The construction of such a questionnaire is comparatively simple and easy because it does not involve setting up conceptual catego

18、ries. However, this kind of questionnaire has very limited value because its results are less generalizable,If the data-analysis is to be on categories , each of which contains a few questionnaire items, you must established either by a top-down approach or by a bottom-up approach. By top-down appro

19、ach you construct conceptual categories based on logical arguments or exsisting theories theories before you design specific items. By bottom-up approach, you simply write down whatever items occur in your mind. You then classify these items into categories,4) Deciding on the content of individual i

20、tem. If a questionnaire is intended to contain several categories , you have to consider whether the items designed cover all the categories and whether each categories contains an adequate number of items. For example,To avoid poorly- worded questions, the following are some of the suggestions made

21、 by Malhotra(1993) and Bernard(1994). Use ordinary words Avoid leading questions Using dual statements: positive and negative Being caution in translation,5) Choosing questions structure. There are two types of questions: open-ended questions or unstructured questions. Open-ended questions refers to

22、 those whose answers are in the respondents own words and are difficult to predict. And these questions are always begin with “wh-”. In fact, this type of question is typically used to collect qualitative data,In closed-questions or structured questions, the researcher provides respondents with a se

23、t of response alternatives. In other words, the respondents have to select one response out of the alternatives provided. They do not have the freedom to use their own words to express their response. The closed-questions may be multiple-choice, dichotomous or a scale,a. Multiple-choice: the researc

24、hers offers a set of answers and respondents is expected to choose one or more alternatives given. b. Dichotomous questions: it has only two alternatives to choose. c. Scales: a scale is continuum upon which a set of alternative choices are placed. For example: Do you guess the meaning of the new wo

25、rds when reading English novels for pleasure,Never Occasionally sometimes often very often 1 2 3 4 5,6) Determining the order of questions. Background information first General questions before specific questions (7) Deciding the format of the questionnaire. Before you print out questionnaire, you n

26、eed to decide the format ,space and the position of the question. Research find that questions place at the top receive more attention than the bottom. (8) Conducting a pilot study to test the questionnaire. A pilot study is one in which a questionnaire is tested on a small sample of respondents to

27、detect and overcome potential problems,5. Scaling techniques,Scaling techniques,Comparative,Noncomparative,Paired comparison,Rank order,Semantic differential,Likert,Paired comparison scaling: as its name suggests, only offers two choices to be the respondent who is asked to compare them according to

28、 some criterion. When you speak English, which of the following two aspects do you pay more attention to? (1) Accuracy (2) Fluency,Rank order scaling Different from paired comparison, rank order scaling expects respondents to order or rank more than two choices with the reference to some certain cri

29、terion. When do you want to pay more efforts into learning English? When I get higher scores on the English test. When I was praised by my English teacher in class. When I obtain a prize in an English speaking competion,Semantic differential scale The semantic differential scale is a five-point or s

30、even-point rating scale with end points associated with bipolar labels that have the opposite semantic meaning. The subject mark the blank that best indicates their views or behaviors. (1) How much do your parents encourage you to study English? Not at all _ ; _ ; _ ; _ ; _ ; _ ; _ ; Very much (2) H

31、ow do you think of Enngish? Interesting _ ;_ : _ ; _ : _ ; _ ; _ ; Boring,One things that has to be emphasized here is that the negative adjectives or phrases sometimes appear at the left side of the scale and sometimes at the right. This controls the tendency of some subjects particularly those wit

32、h very positive or very negative attitudes, making the right- or left-hand side without reading the labels,The subjects were asked to evaluate their reasons for learning English on a four point-scale. 1 = not important 2 = somewhat important 3 = important 4 = very important It is called a Likert sca

33、le and it was first introduced by Rensis Likert,To conduct the analysis, each response is assigned a numerical score, ranging from 1 to 4. When we use this approach to determine the total score for each subject, it is important to use a consistent scoring procedure so that a high (or low) score cons

34、istently reflects a favorable response. This requires that the numerical value assigned to the negative statement by the subjects be scored by reversing the value. Note for a negative statement, an agreement reflects an unfavorable response, whereas for a positive statement, reflects a favorable res

35、ponse. Accordingly, a “strong disagree” to a negative statement would both receive the same score,6. Selecting subjects,You cannot study everyone anywhere doing everything ” Miles and Huberman,1994: 27 The common practice is to select a sample from the population to study, hoping the findings from t

36、he sample can be applied to the whole,6.1 Random sampling techniques There are three basic random sampling techniques: simple random sampling, systematic random sampling and stratified random sampling. (1) Simple random sampling It is the basic for the other two. By a Simple random sampling, each un

37、it is numbered from 1 to N (the size of population). Next, a table of random number is used to select n items to the sample,2) systematic random sampling If the population is large but the intended sample size is small, simple random sampling is not suitable because the sample selected might not be

38、evenly distributed among the population. It can make sure that the subjects selected are evenly spread out among the population, and we need to know the interval. Interval = the total number of the population/the size of the sample If the interval happens to be number with a decimal, you need to rou

39、nd it off into a whole number. 29.4-29 29.5-30,Suppose that the total number of a population consists of 900 students and you would like to select 30 students out of them as a random sample. The interval is 900/30=30 The first number is 915900 (No) The second 674 (OK) The third 674+30 674,704,734,764,794,824,854,884,914-900,3) Stratified random sampling For example: 900 students, 30 female. a. You want to get a sample of 30 students in which