Five questionnaire study are summarized analysis of ideas

Questionnaire as one of the most commonly used time-saving, things can be more comprehensive system of survey methods, both are acclaimed in their daily work or academic research. But once the survey is not so easy to complete, with the questionnaires have been collected, a large amount of data waiting to be sorted, statistics and analysis. So much data before the start of it from where to start? You should analyze what specific way to do that?

For five analytes ideas, usually applied to social sciences, including business administration, tourism management, marketing, and psychology major, education, teacher, language and other majors. Description of these five ideas for specific analytical framework is as follows: 

     

 

 

1, the relationship between class Effect

 

 

 

Affect the relationship the most common type of research, the study affect the relationship, often to draw up model framework, a framework for the overall study presentation structure ideas, the research framework as the core. Such analysis of ideas pour more important than the relationship, and the relationship is the study of the impact. Effects of various factors such as the relationship between employees pay satisfaction, employee turnover tend to affect the relationship between research, consumer repeat purchase intention impact studies relations. Such analysis framework for thinking questions most items should scale problem, a small amount of a non-title item scale problem. Psychology, management class, tourist class, marketing and other professional use higher frequencies framework of this analysis thinking, thinking this analysis framework applicable to all readers.

 

Specific research content:

1. First, the data collection for basic frequency analysis, such as statistical distribution of gender, age, education of such a system.
2. If the characteristics of the case study involved a sample, such as basic behaviors, attitudes or cognitive relevance, frequency of use can also be pooled analysis further understand clearly the characteristics of the sample cases.
3. Studying the relationship between impact questionnaire usually involves a lot of scale questions, if the scale problem should be divided into specific number of dimensions, not entirely sure, then you can use factor analysis was concentrated, leads to several dimensions (factors) and find the dimensions of the problem situation correspondence between items. (Note: a plurality of dimensions represented by a heading, the heading would like to summarize the plurality into a whole, in this case need SPSSAU "generated variable" and "average" can function)
4. After the reliability of the data, whether there is reliability, is the most basic, general background on the basic characteristics of the sample, because first of all have to know is what kind of answer to question sample population group. At the same time reliability for scale study only type of data, such as gender can not be analyzed against the background of an item of information like age.
5. Aside from credible data, but also we need to study the scale items have reliability problems. Prior reliable data, reanalysis effective, it is a common configuration, validity analysis and reliability analysis are also used interchangeably position.
6. Reliable data, and then scale research is valid, then the need for specific dimensions (Scale title items, etc.) are described analysis, the basic attitude of the case study sample population for the scale items.
7. Inventory items complete title, after a descriptive analysis of each dimension, then use correlation analysis to study the relationship between the cases, regression analysis to prepare.
8. Under the data associated with the premise, and then return to study the impact relations have meaning. After correlation analysis therefore needs to be placed regression analysis. And require the use of regression analysis to test hypotheses normally.
9. There may also need to compare different groups of people, such as sex, age and other different groups, they question the attitude of discrepancy scale items, which generally can be analyzed using analysis of variance, or T inspection. If people want to study different backgrounds (such as gender, age) for the differences in the behavior of the sample, the proposed cross chi-square analysis, etc. can be used, and if the cross-analysis of multiple-choice questions involved, but also need to select the corresponding method.

 

2, the status quo policy class research

 

 

 

Status quo policy is a very common type of research, analysis of ideas in the comparative study and dumping heavy and basic attitudes, and differences, to understand the basic cognitive groups, attitudes, views, or opinions behavior. Class research status quo policy is usually more items are non-Scale questions in the case, the core framework of such studies is that "group." The first thing is the "group"; for example, there are 30 questions, 30 questions here that this item can be summarized in a few aspects of it were? Such as the basic background, knowledge, attitudes, behavior, reasons five aspects. The second is that the "packet", respectively, as part of analysis. The third thing is a cross between a group title and group title entry items. Sociology class, and other related professional media use this analytical framework related research, as well as a questionnaire study business ideas is more suitable for this type of analysis framework.

 

Specific research content:

1. First, the data collection for basic frequency analysis, such as statistical distribution of gender, age, education and so on of how
2. For each "packet" is analyzed, each "group" as a portion for analysis.
3. Differences relationship: for example, the basic background are the "cognitive", the difference on the "attitude", "behavior", "reason" (usually using a cross-analysis, of course, when it comes to multiple-choice questions, you should select the corresponding multiple choice should cROSS used)
4. If want to study affect the relationship, such as research "cognitive" (X), "Attitude" (X), respectively, for "behavior" (Y) effect relationship, then you can consider using a binary logit regression analysis. But it requires special attention to the fact that: "Behavior" (Y) must have only two options, such as willing and unwilling, to buy and not to buy, and you must use the numbers 1 and 0, respectively. For example, studies the impact of online shopping attitude "whether to buy purchasing products", should be willing to be a number 1, I do not want to use number 0.

 

• NOTE: If you have questions Scale questionnaire, you can also use reliability analysis, variance analysis and other methods.
• NOTE: Do not use non-Scale Method data reliability, validity and so on.

 

3, the adjustment / intermediary class Study

 

 

 

Regulating / intermediation framework, "Effects of Relationship" that is similar to the first category. However, such a framework is more common in academic research, and thus may be more a mediating role, or regulation of the two parts, the base and the rest of "Relationship influence" essentially similar.

Such analysis dumping heavy thinking intermediary role, or to study the role of regulation. For example, when employees work-life balance study the impact on turnover intention, job satisfaction whether mediates. When product quality willingness to spread awareness on the impact of word of mouth, whether different income levels play a regulatory role. The first analysis framework similar ideas, ideas such analyzes in the framework of a large number of questions on the questionnaire design items are scale problem, only a small part of the problem is non-Scale item title item. Business administration, marketing, psychology, education and other professional prefer to use the idea of such an analysis framework, the analysis of ideas framework more suitable for a certain statistical basis of the reader.

Note that, in the academic research studies are very cautious intermediary role or regulation, so if these two types of research needs, need to be described in detail in the literature review clearly intermediary or regulation relationship before doing any analysis (that is, first theoretical basis, and then use the research methods of verification), otherwise there will be a logical problem.

 

4, experimental class Differences

 

 

 

Experimental studies class, often referred to as class research "scene" and so on. Such frameworks in market research are relatively common. Relatively speaking, the core of such studies is that the interactions, and often are using a scale questionnaire class title.

Such analysis of ideas for using the experimental approach to the relationship between the questionnaire and the form of research, questionnaire design will typically include different profiles. Music department stores such as the impact of consumption will stimulate research, usually a questionnaire which will have different scenarios, such as background music or no background music, or there are also different types of background music. Such analysis mentality framework emphasizes differences under different scenarios or different experimental conditions, there will be more entries scale problem usually. Marketing, psychology, media and other related professionals will use this analytical framework for various studies.

 

5, sample class research cluster

 

 

 

Analysis of ideas tend to cluster sample study sample "classified", that is, the sample population should be divided into several categories ; after sub-categories, then the difference is usually definitely need to compare different categories of people, such as different types of groups in attitude, behavior differences and so on. It should be noted that the design of the questionnaire items may question the scale used for clustering samples, as well as such non-Scale questionnaire will have more problems items, used to understand the characteristics of the sample cases. Social Studies, Marketing or related professional often use this analytical framework to study the idea of this analysis framework more suitable for a certain statistical basis of the reader.

Specific research content:

1. First, the basic characteristics of the sample case study data (such as gender, age, education, etc.) for analysis
2. If there are groups involved in sample characteristics, behavior, attitude or related items can be part of a separate analysis
3. If the data does not know the scale research into several dimensions, such as a 20 gauge question, should be divided into several dimensions are not determined at this time may be performed using factor analysis. Find should be divided into several dimensions (factors), and the correspondence between items and dimensions of the problem situation.
4. Next, the scale items reliability and validity analysis (Note: the dimensions have been derived factor analysis item corresponding relationship problems, this time validity has been described, sometimes may forgo validity analysis, from the recommended content integrity into )
5. After completion of the factor analysis has been confirmed several dimensions, it may be several dimensions of this cluster, to give groups of several categories, and the binding characteristics of each type of group, named category for each cluster. (Note: a dimension represented by a plurality of question items, to summarize the plurality of title key as a whole, in this case need SPSSAU "generated variable" and "average" can function)
6. After the obtained cluster categories, then the differences need to compare different types of groups; include, for example differences in the "features", "act" or the "attitude" of. Easy to combine different groups provide different advice measures.

 

Understand the idea of analysis for each type of questionnaire after their own design corresponds to find the type of questionnaire belongs, is analyzed to follow the steps. Specific analysis may SPSSAU in progress, a key generating drag intelligent analysis results.