JP2015064666A - Questionnaire analysis system and program thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To tabulate questionnaire answer data, to analyze and illustrate to what question item attentions are paid by answerers and how tendencies of tabulation results are possible to deviate.SOLUTION: In a questionnaire analysis system, a dependent variable is calculated and considered as an importance value by performing multiple regression analysis for deriving a multiple regression formula for converting choices of answer data into first numerical data according to weighting, using any one or a plurality of question items of a questionnaire as total evaluation question items, using the first numerical data corresponding to the choices of the total evaluation question items as an objective variable, and using the first numerical data corresponding to choices of the respective question items except the total evaluation question items as an explanatory variable. In addition, an average value of the first numerical data is calculated and considered as a satisfaction value. An attention degree value is further calculated by tabulating the number of keywords to be included in comment character strings of the answer data.

Description

  The present invention relates to a questionnaire analysis system for collecting and analyzing questionnaire response contents and a program thereof.

  In general, in order to objectively know improvement items of facilities and equipment, a questionnaire is conducted for a plurality of users who actually use the facilities and equipment. The questionnaire is conducted by answering by facility and equipment about the satisfaction and comfort of using the facility and equipment. Then, by summing up the answer results of the questionnaire, it is possible to know the facilities and equipment with low overall satisfaction and usability.

  The purpose of such a questionnaire is to improve the user's satisfaction with the facility and equipment by objectively knowing the improvement items of the facility and equipment and improving the facility and equipment. However, most companies have limited budgets for renovating facilities and equipment.

  Therefore, once the answers to the questionnaire are tabulated, the tabulated data is analyzed, and the facilities and equipment that have a low overall evaluation such as user satisfaction and usability are preferentially improved, and the user's satisfaction with the facilities and equipment is improved. Efforts are made to improve.

  Therefore, it is a very important matter to accurately analyze the aggregated data of questionnaire responses and to prioritize which facilities and equipment to improve. The questionnaire response data can be aggregated by simply counting the number of options that have been answered, by calculating the average value by weighting the response items of the options, or by statistical methods such as multiple regression analysis. A counting method is used.

  As a method of totaling which question items are considered to be comprehensively important among the question items, for example, in Patent Document 1, questionnaire responses are totaled according to environmental components and usage scenes, and are implemented with priority. A method is disclosed that allows the user to visually grasp the countermeasure to be taken.

  In addition, as a method for displaying the total data, for example, in Patent Document 2, the normalized evaluation value to be investigated is displayed for each of a plurality of items on a four-quadrant graph, whereby relative superiority or inferiority between items is displayed. A method for enabling comparison and displaying the analysis result in a format that is easy for the user to grasp is disclosed.

  Furthermore, in Patent Document 3, the content of the questionnaire answered in text is analyzed by text mining, and the information indicating the deviation of the tendency of the options estimated from the text is obtained by associating the selective answer type question with the text. The technology to create is disclosed.

JP 2009-129124 A JP 2004-127308 A JP 2005-44190 A

  By using techniques such as those disclosed in the prior art documents, it is possible to calculate the answers to the questionnaire according to the purpose and know what tendency the total results have. For example, by combining the techniques disclosed in Patent Documents 1 to 3, the questionnaire response items are aggregated by item, and the aggregation results are displayed on a graph, etc. It is possible to make a judgment as to which item should be dealt with preferentially.

  However, the conventional technology aggregates the data that are answered in the form of choices for the questionnaire question items. For this reason, even if contents that contradict the answer selected by the option are described in the free comment column of the questionnaire, the contents of the free comments are not reflected in the aggregated data.

  For example, if you want to reply that you are satisfied with a certain question item, but are dissatisfied with a part, select “satisfied” as the option, and include some dissatisfaction with the free comment. This is a case where the content is answered in text.

  In such a case, if the ratio of “satisfied” to the options is large, it is considered that the question item is evaluated as being satisfactory as the final comprehensive evaluation. If the final comprehensive evaluation is “satisfied”, even if the free comment contains a dissatisfied sentence for the question item, the written response will not be reflected in the comprehensive evaluation. Aggregate results may not match the actual status of the answers.

  In addition, even if there is an important opinion in such free comments, it may be buried in the total tabulation and may be excluded from consideration. I can't know exactly what I want. In such a case, even if the total evaluation of the total result is “satisfied”, if the same questionnaire is performed again, the total evaluation of the total result may be deteriorated.

  As a method of analyzing a free comment sentence and associating it with a selective answer type question to create information indicating a shift in the tendency of options, a method as disclosed in Patent Document 3 is disclosed. However, the method according to Patent Document 3 is intended to grasp a shift in tendency between the choices of individual question items and the answers to free comments.

  For this reason, it is possible to recognize a difference in tendency between a choice-type answer and a descriptive-type answer to the same question. It is difficult to use it to figure out if there is.

  The present invention has been made in order to solve the above-mentioned problems, and the purpose of the present invention is to determine which question items are focused on the respondents when displaying the trend of the total data of the questionnaire response data, and the total results The purpose is to provide a questionnaire analysis system and its program that allow accurate questionnaire response data to be aggregated and analyzed by illustrating how the trend of the survey may change .

  The questionnaire analysis apparatus according to the present invention for solving the above-described problem is to aggregate the contents of responses of a plurality of respondents to a question item of a questionnaire composed of a plurality of questions, and evaluate the respondents of the evaluation items corresponding to the question items. A questionnaire analysis system for analyzing the answer data, the answer data input accepting section for accepting the answer data input of the choice and comment character string respectively answered by a plurality of question items of the questionnaire, and the answer data input accepting section. The answer data storage unit for storing the selected option and the comment character string, the keyword storage unit for storing a plurality of keyword information extracted from the comment character string, and the options stored in the answer data storage unit according to weighting Any one of the weighting correction unit for converting to the first numerical data and the question item of the questionnaire, or The number is the comprehensive evaluation question item, the first numerical data corresponding to the choice of the comprehensive evaluation question item is the objective variable, and the first numerical data corresponding to the choice of each question item other than the comprehensive evaluation question item is the explanatory variable By calculating the multiple regression analysis to derive the multiple regression equation, the dependent variable of each question item other than the comprehensive evaluation question item is calculated respectively, and the calculated dependent variable is evaluated corresponding to each question item other than the comprehensive evaluation question item An importance value calculation unit for calculating the first importance value of the item, and an average for calculating the first average value from the first numerical data corresponding to the choices of each question item other than the comprehensive evaluation question item of the questionnaire By calculating the number of keywords stored in the keyword storage unit included in the comment character string stored in the value calculation unit and the answer data storage unit, The attention result value calculation unit for calculating the attention value of the evaluation item and the analysis result for storing the calculated first importance value, first average value, and attention value for each evaluation item corresponding to the question item And a storage unit.

  By adopting such a configuration, it is possible to calculate an average value of options by digitizing questionnaire response data according to the weight of the options. In addition, by calculating by the multiple regression analysis method, it is possible to calculate the importance value such as which questionnaire item is relatively emphasized by the respondent of the questionnaire. Furthermore, by collecting the keyword of the comment character string and the questionnaire items in association with each other, it is possible to also calculate the attention value such as which questionnaire item is attracting attention. By summing up the average value, importance value, and attention value together, an indicator of how much the questionnaire's question items are averaged from the average value and which questionnaire item is calculated from the importance value It is possible to simultaneously obtain an index such as whether or not the questionnaire item is considered important, and an index such as which questionnaire item is attracting attention from the attention value. As a result, it is possible to more reliably determine which evaluation item corresponding to which questionnaire item should be preferentially supported.

  The questionnaire analysis apparatus according to one aspect of the present invention is the questionnaire analysis system, wherein the weight correction unit further includes a keyword stored in the keyword storage unit in a comment character string, The first numerical data of the options answered to the question items related to the keyword is converted into the second numerical data according to the weight of the comment, and the importance value calculation unit further selects the comprehensive evaluation question item options. By performing multiple regression analysis to derive a multiple regression equation using the second numerical data corresponding to the objective variable and the second numerical data corresponding to the choice of each question item other than the comprehensive evaluation question item as an explanatory variable, Dependent variables for each question item other than the comprehensive evaluation question item are calculated, and the calculated dependent variable is an evaluation term corresponding to each question item other than the comprehensive evaluation question item. The average value calculating unit further calculates a second average value from second numerical data corresponding to the choices of each question item other than the comprehensive evaluation question item of the questionnaire, The analysis result storage unit is further an analysis result storage unit that stores the calculated second importance value and second average value for each evaluation item corresponding to the question item.

  By adopting such a configuration, it is possible to obtain an index as to how the satisfaction value or importance value based on the average value may change based on the contents described in the comment character string. It becomes like this. As a result, judgments regarding which evaluation items should be preferentially handled, such as preferentially responding to evaluation items corresponding to questionnaire items that are highly likely to change significantly, are made in a more realistic manner. become able to.

  The questionnaire analysis apparatus according to one aspect of the present invention is the questionnaire analysis system, wherein each evaluation item is stored in the analysis result storage unit on a four-quadrant graph having the importance value and the average value as axes. It is characterized by comprising an analysis graph creating unit that creates graph image data plotted with a size according to the ratio of the attention level of each evaluation item based on the first importance value and the first average value. To do.

  By adopting such a configuration, we can make comparative comparisons more appropriately by visually expressing values such as satisfaction values, importance values, and attention values based on the average value of multiple questionnaire items on the graph. Will be able to.

  The questionnaire analysis apparatus according to one aspect of the present invention is the questionnaire analysis system, wherein the analysis graph creation unit further stores each evaluation item in the analysis result storage unit on the four-quadrant graph. Based on the importance value of 2 and the second average value, graph image data plotted in a shape that can recognize the change between the first importance value and the plot of each evaluation item based on the first average value is created It is characterized by doing.

  By adopting such a configuration, not only the values such as the satisfaction value, importance value, and attention value based on the average value of multiple questionnaire items are visually represented on the graph, but the change prediction is also graphed. By expressing the above, it becomes possible to perform a more reliable comparative study.

  Moreover, the questionnaire analysis program of this invention is a questionnaire analysis program characterized by operating a computer as the said questionnaire analysis system.

  According to the present invention, by calculating three indicators such as a satisfaction value, an importance value, and an attention value based on the average value from the answer data of the questionnaire, which evaluation item of the evaluation item corresponding to the question item of the questionnaire is determined. It can be compared whether it is relatively satisfied with other evaluation items, is considered important, or is attracting attention. Further, visual judgment can be made by displaying the index on the four-quadrant graph.

  Further, by displaying the satisfaction value and importance value together in a four-quadrant graph, for example, an evaluation item with a low degree of satisfaction although the importance value is high, the satisfaction value and the importance value are displayed. It becomes possible to compare evaluation items in combination.

  In addition, by separately calculating the attention level value, it is possible to find an evaluation item that is not important but has received attention. Furthermore, by changing the weighted numerical value of the answer data options based on the comment and calculating the satisfaction value and importance value based on the first numerical data and the second numerical data, the satisfaction level and the important value are calculated. The degree of change can be predicted.

It is a conceptual diagram of questionnaire embodiment. It is the figure showing the example of the implementation format of the question and answer of a questionnaire. It is a block diagram of a questionnaire analysis system. It is an example of the data structure of questionnaire data. It is an example of the data structure of the answer data of a questionnaire. It is an example of the data structure which memorize | stores the answer data of a questionnaire as 1st numerical data. It is an example of the data structure which memorize | stores the answer data of a questionnaire as 2nd numerical data. It is an example of the data structure of keyword information. It is an example of the data structure which totaled the number of cases of answer data. It is an example of the data structure which totaled the number of response data according to attribute. It is a block diagram of an answer data analysis part. It is an example of the data structure which memorize | stores an analysis result. It is an example of the data structure which memorize | stores an analysis result. It is an example of the figure which represented the number of respondents according to attribute as tabular data. It is an example of the figure which represented the ratio of the number of respondents according to attribute as a pie chart. It is an example of the 4-quadrant graph which displayed the analysis result. It is an example of the 4-quadrant graph which displayed the analysis result. It is an example of the 4-quadrant graph which displayed the analysis result. It is an example of an analysis report. It is an overall process flowchart of a questionnaire analysis system. It is a process flowchart which calculates 2nd numerical data. It is a process flowchart which calculates an attention degree.

  Hereinafter, embodiments of a questionnaire analysis system and a program thereof disclosed in the present application will be described with reference to the drawings. In the following example, in order to improve the office environment, a questionnaire for employees working in the office is conducted, and the equipment in the office environment that the employee wants to improve is determined. An explanation will be given by taking a response analysis of a questionnaire for investigation. However, the questionnaire analysis system of the present invention can be applied not only to response analysis of a questionnaire for improving the office environment, but also to response analysis of any questionnaire.

  FIG. 1 is a conceptual diagram showing how a questionnaire is implemented and how a questionnaire analysis system collects questionnaire responses and outputs analysis results in this embodiment. As shown in FIG. 2, the questionnaire format in this embodiment is a questionnaire format consisting of a question for selecting an option for a question item and answering the questionnaire, and a comment answer freely entered. Is assumed.

  As an embodiment of the questionnaire, (1) questionnaire questionnaire items are entered on paper, and the questionnaire is collected by filling in answers on paper. (2) The questionnaire execution system is operated on the WEB server, and the answers to the questionnaire are input from a browser or the like. (3) Look at the questionnaire question items, such as paper, and collect the answers as electronic data. Such a questionnaire embodiment is conceivable.

  For example, (1) As a method for collecting questionnaires by filling out questionnaire question items on paper and filling in answers on paper, the answers to the questionnaire question items printed on the questionnaire answer sheet 400 of paper data are in mark sheet format. A plurality of questionnaire response sheets 400 are collected by filling in a questionnaire response number. The collected questionnaire response sheet 400 is mechanically read an answer if it is written in the form of a mark sheet, or is taken into the questionnaire analysis system 100 as answer data. If the answer number is to be entered, the person in charge of questionnaire collection reads the answer number, converts the answer number and answer contents into electronic data, and imports them into the questionnaire analysis system 100.

  (2) As a method of operating a questionnaire execution system on a WEB server and having a questionnaire response input from a browser or the like, the questionnaire execution system is operated on the questionnaire execution server 200. Connect to the questionnaire execution server 200 from 300 WEB browsers, etc., and display the questionnaire on the WEB browser. Then, a questionnaire response is input on the WEB browser, and the response is transmitted to the questionnaire implementation server 200. The questionnaire execution server 200 aggregates input questionnaire response data and transmits it to the questionnaire analysis system 100.

  (3) As a method of collecting and collecting answers as electronic data by looking at questionnaire question items such as paper, the questionnaire respondents create only questionnaire response data as questionnaire response sheet 500 as electronic data For example, questionnaire response data is collected by transmitting the response electronic data of the questionnaire response sheet 500 to the questionnaire analysis system 100 by e-mail.

  The questionnaire analysis system 100 collects questionnaire response data by the method as described above, and after tabulating and analyzing, the tabulation and analysis results are printed on paper as a questionnaire analysis report 600 or displayed on the screen. Do things such as.

  The person who conducted the questionnaire confirms the questionnaire analysis report 600 and takes measures to be taken according to the questionnaire implementation result. For example, in the case of a questionnaire for improving the office environment, the questionnaire analysis report 600 is used as a material for determining which equipment in the office environment should be renovated with priority.

  The above is a brief description of the questionnaire implementation method, the questionnaire data collection method, and the output of the questionnaire analysis report 600. Next, it will be described in detail what questionnaire response data the questionnaire analysis system 100 collects and how to analyze it.

  FIG. 3 illustrates the configuration of the questionnaire analysis system. The questionnaire analysis system includes a questionnaire storage unit 110, an answer data input reception unit 101, an answer data storage unit 111, an answer data totaling unit 102, an answer data analysis unit 103, a keyword storage unit 112, an aggregate data storage unit 113, and an answer data totaling unit. 104, an analysis result storage unit 114, an analysis result output unit 105, and an analysis graph creation unit 106. Hereinafter, each component will be described.

  First, the questionnaire storage unit 110 will be described. The questionnaire storage unit 110 is a storage unit that stores questionnaire data of questionnaire items to be analyzed and answer options. The questionnaire storage unit 110 stores questionnaire question items, options, and weight values compared with other options. FIG. 4 shows an example of the data structure of the questionnaire data stored in the questionnaire storage unit 110 for conducting an investigation of the office environment. A questionnaire as shown in FIG. 2 is stored as questionnaire data as a data structure as shown in FIG.

  The questionnaire data stores information of “identifier”, “question type”, “question item”, and “answer selection item (weight)”. In the “identifier”, information on whether the question item is a question regarding the attribute of the respondent or a normal question and a question number are stored. The attribute is a question for confirming what kind of position or preference the respondent is, such as a question about gender or work style. On the other hand, a normal question is a question necessary for obtaining necessary information for the purpose of a questionnaire. In “Question type”, a question type such as what question the question is is stored. The question type is also handled as an evaluation item for the question item. In “Question Item”, a question character string asking a question is stored, and in “Answer Selection Item (Weight)”, a list of options for the question item and a weight value of each option are recorded. For example, in FIG. 4, the higher the satisfaction level of the option, the higher the weighting of the numerical value.

  The information stored in the questionnaire storage unit 110 includes, for example, aggregation by attribute when the response data totaling unit 102 or the response data analysis unit 103 performs totalization or analysis of the response data, or performs analysis using the weight value of the option. Used as information.

  Next, the response data input receiving unit 101 is a processing unit that receives input of questionnaire response data 700 to be analyzed. The answer data input accepting unit 101 accepts, as the questionnaire answer data 700, for example, the answer data converted into text data in CSV format as input data. As described with reference to FIG. 1, since there are several methods in the embodiment of the questionnaire, it is not always necessary to accept the input of the questionnaire response data as text data in CSV format, and the response data is directly received from the questionnaire execution server. Alternatively, the answer data may be directly input and received via the input interface.

  When the answer data input receiving unit 101 receives the input of the answer data, the input answer data is recorded in the answer data storage unit. FIG. 5 shows an example of the data structure of the answer data stored in the answer data storage unit 111. In FIG. 5, information of “answer date / time”, “respondent ID”, “answer content of options by identifier”, and “comment” is stored as answer data.

  As these items, the items of the answer data input by the answer data input receiving unit 101 are stored as they are. “Reply date” is the date and time when the respondent entered the questionnaire response, “Respondent ID” is the identifier of the respondent who answered the questionnaire, and “Reply content of options by identifier” is a question for each identifier The contents of the answer of the respondent's option for the answer are stored in “Comments”, respectively. Basically, “answer contents by identifier” stores the answer data for the question item of each identifier distinguished by “identifier” in the questionnaire storage unit 110 by the number of identifiers.

  Next, the weight correction unit 102 will be described. The weighting correction unit 102 digitizes the answer data of the answer data stored in the answer data storage unit 111 in accordance with the weighting of each option stored in the questionnaire storage unit 110. Data obtained by digitizing the options according to the weights is stored in the aggregate data storage unit 113 as first numerical data. Further, the weight correction unit 102 corrects data in which each option of the answer data is digitized based on the keyword information stored in the keyword storage unit 112. The digitized data obtained by correcting the digitized data based on the keyword information is stored in the total data storage unit 113 as second numeric data. How to correct the weighting of the numerical value of each option based on the keyword information will be described later.

  FIG. 6 shows an example of the data structure of the answer data in which the choices stored in the total data storage unit 113 are quantified. Basically, the item of the data structure of the answer data stored in the total data storage unit 113 is the same as the example of the data structure of the answer data stored in the answer data storage unit 111, but the item of the answer contents by option Stores numerical values corresponding to the respective options. These numerical values are handled as the first numerical data.

  FIG. 7 shows a data structure in which the numerical values of the answer data in which the options stored in the total data storage unit 113 are digitized are stored as numerical values corrected based on the keyword information stored in the keyword storage unit 112. An example is shown. This data structure is also basically the same as the example of the data structure of the answer data stored in the answer data storage unit 111, and the items stored are the same, and the answer content items for each option include numerical values corresponding to each option. Each of the values obtained by correcting the converted values based on the keyword information is stored. These numerical values are handled as second numerical data.

  The second numerical data calculation process will be described. FIG. 8 shows an example of the data structure of the keyword information stored in the keyword storage unit 112. As the keyword information, a plurality of items of data “question type”, “keyword”, and “keyword correction value” are stored. “Question type” stores information regarding to which question type the corresponding keyword is related, “Keyword” represents a keyword character string extracted from the entered comment, and “Keyword correction value” corresponds The negative value that the keyword means is a keyword relative to other keywords. If the correction value for a keyword that does not evoke a special emotion image is 0, a larger value indicates a keyword with a relatively positive meaning, and a negative value indicates a relatively negative value. Represents a meaningful keyword.

  The weighting correction unit 102 determines whether the comment entry field of the questionnaire response includes a character that matches the keyword of the keyword information stored in the keyword storage unit 112. If the keyword is included in the comment entry field, The second numerical data is corrected by adding the value of the keyword correction value to the numerical value of the answer value of the answer data of the question of the question type corresponding to the keyword, thereby correcting the numerical value of the answer data. Is calculated.

  Next, the answer data totaling unit 104 will be described. The answer data totaling unit 104 totals the answer data stored in the total data storage unit 113 and stores it in the total data storage unit 113. The total item counts the total number of answers and the number of answers that have been answered for each question type. The answer data is aggregated for the first numerical data.

  The response data is totaled by total response data and by identifier attribute. Aggregation is performed for each question type corresponding to each question, which is the answer selection item that is the answer to the question of the identifier question. The answer selection items are tabulated for each option corresponding to each question stored in the questionnaire storage unit 110.

  FIG. 9 shows an example of the data structure of data obtained by counting the total number of response data. In FIG. 9, “total number of answers”, “answer items”, “weight”, and “question type” are included as total items. The “question type” includes items of a plurality of question types corresponding to the question types in the questionnaire storage unit 110. As the “number of responses”, numerical values of the number of responses of the entire response data are tabulated by the answer data totaling unit 104 and recorded. In “answer item” and “weight”, options corresponding to each question stored in the questionnaire storage unit 110 are stored as items, and in “question type”, stored in the questionnaire storage unit 110. A question type item corresponding to each question is stored.

  In the answer data totaling unit 104, the number of answers to each question is counted for each option, and the number is recorded for each question type. When the correspondence between the question and the question type is a one-to-one relationship, the number of answers for each option that is totaled in the question type is a total value of the number of answers to the question. When multiple questions are associated with a single question type, the number of answers for each option compiled in the question type is the number of answers for the associated question for each question. This is the sum of the values obtained.

  The answer data totaling unit 104 calculates the answer data not only for the entire answer data but also for each attribute of the identifier. FIG. 10 shows an example of a data structure in which answer data is tabulated by identifier attribute. FIG. 10 shows an example of the data structure of the result of the same aggregation as the total number of response data items for each attribute. In “attribute”, an identifier stored in the questionnaire storage unit 110 is an item of an answer option for a question item having an attribute. For each selection item that is answered to the question about the attribute, the number of answer data for the question with the questioned identifier is aggregated. Therefore, the “number of cases” stores the number of cases in which the question questions regarding the attribute are answered.

  The above is an explanation of how the aggregated data aggregated by the answer data aggregating unit 104 and the answer data aggregating unit 104 is stored in the aggregated data storage unit 113.

  Next, the answer data analysis unit 103 will be described. As shown in FIG. 11, the answer data analysis unit 103 further includes an importance value calculation unit 1031, an average value calculation unit 1032, and an attention value calculation unit 1033. Each processing unit of the answer data analysis unit 103 reads each tabulated data from the tabulated data storage unit 113 and calculates and calculates various index values by analyzing and calculating the answer data.

  First, the importance value calculation unit 1031 will be described. The importance value 1031 is obtained by performing multiple regression analysis using the numerical data of options for each question item in FIGS. 6 and 7, so that the answer of the question type corresponding to each question item is the total of the answer data of the entire questionnaire. The importance value of each question type is calculated by calculating a coefficient such as how much influence is exerted on the general evaluation.

  Here, the digitized data corresponding to the answer data choices for one or more of the questionnaire questions is used as the objective variable, and the digitized data corresponding to the answer data choices for the other questions is supported for the questions. As an explanatory variable for the question type to be calculated, a coefficient of each question type can be calculated by creating a multiple regression analysis formula based on the answer data of a plurality of respondents and performing multiple regression analysis. This coefficient is recorded in the analysis result storage unit 114 as an importance value.

  More specifically, among the question data stored in the questionnaire storage unit 110, numerical data of choices of answer data of a question whose question type is “overall evaluation” is used as an objective variable, and other question types Let the dependent data be the digitized data of the answer data of the question. In the question data shown in FIG. 4, the question type is the comprehensive evaluation, and the question data whose identifier is question 1 is the comprehensive evaluation. Accordingly, the data obtained by quantifying the answer data choices with the identifier of question 1 is the objective variable, and the data obtained by quantifying the answer data choices of the questions with the identifiers of question 2 through question 10 is the dependent variable.

  Here, a general multiple regression analysis formula is expressed as a multiple regression analysis formula of “target variable = dependent variable 1 × coefficient 1 + dependent variable 2 × coefficient 2 + dependent variable 3 × coefficient 3+. The importance value calculation unit 1031 uses this formula and the digitized data of the answer data of the questions of each question type, “total evaluation value = noise value × noise importance value + air environment value × air environment” The multiple regression analysis formula is established as “importance value of wiring + numerical value of wiring and flow line × importance value of wiring and flow line ...”, and each of the multiple regression analysis formulas shown in FIG. By applying the first digitized data of the respondent, each first importance value corresponding to each question type is calculated. The calculated first importance values are recorded in the analysis result storage unit 114. Similarly, by applying the second digitized data as shown in FIG. 7, the second importance value corresponding to each question type is also calculated. Each calculated importance value is recorded in the analysis result storage unit 114 in a data structure as shown in FIGS.

  Next, the average value calculation unit 1032 will be described. The average value calculation unit 1032 calculates a value obtained by averaging the weight values of the choices of answer data of corresponding question items for each question type in FIG. The average value is calculated simply by adding all the values (first numerical data) of the answer data choices for each question by weight and dividing by the number of answer data. Calculate the average value of numerical data. When there are a plurality of questions corresponding to the question type, the average value of each question is calculated, the calculated average values are added together, and the average value of the question type is calculated by dividing by the number of questions.

  For example, the first average values are calculated by using the first digitized data of the answer data for each question as shown in FIG. The calculated first average values are recorded in the analysis result storage unit 114. Similarly, the second average value is calculated by using the second digitized data as shown in FIG. Each calculated average value is recorded in the analysis result storage unit 114 in a data structure as shown in FIGS.

  The above average value is calculated for all response data, but when calculating the average value using the first numerical data, not only the average value of the entire response data, The average value for each question type for each attribute can also be calculated by using the total value of the answer data choices for each attribute as tabulated in FIG.

  Assuming that the average value is weighted with a larger value as the answer to the question item is affirmative, the average value can be treated as a satisfaction value, and the higher the average value, It can be considered that the satisfaction value is high.

  Next, the attention value calculation unit 1033 will be described. The attention level value calculation unit 1033 calculates the attention level value depending on how many keywords stored in the keyword storage unit 112 are included in the comment of the answer data as illustrated in FIG. 6 stored in the total data storage unit 113. calculate. Specifically, for each keyword stored in the keyword storage unit 112, it is determined whether or not the keyword is included in the comment of the answer data stored in the total data storage unit 113. If a keyword is included, the attention level value of the question type corresponding to the keyword is added by the number of keywords included. For example, if the keyword “noisy” appears three times in the comment of the answer data, the attention value of the question type “noise” corresponding to the keyword “noisy” is incremented by 3.

  Similarly, the number of times of appearing in the comment of the answer data for all the keywords stored in the keyword storage unit 112 is totaled, and the attention value is calculated for each question type. The calculated attention level value is recorded in the analysis result storage unit 114 in a data structure as shown in FIG.

  The above is the description of each processing unit of the answer data analysis unit 103. Next, the analysis result output unit 105 and the analysis graph creation unit 106 will be described. The analysis result output unit 105 and the analysis graph creation unit 106 format the analysis data stored in the analysis result storage unit 114 and the total data stored in the total data storage unit 113 and output the result as a questionnaire analysis report 600. .

  The questionnaire analysis report 600 may display data representing a questionnaire analysis result as a table or a graph, for example, on a display monitor, or may be printed as a paper report as a form report. May be. For example, the analysis result output unit 105 outputs the number of respondents by attribute stored in the total data storage unit 113 in a tabular format. FIG. 14 shows an example in which the number of respondents by attribute is output as tabular data. The analysis result output unit 105 outputs not only the number of respondents by attribute but also tabulation data and analysis data stored in the tabulation data storage unit 113 and the analysis result storage unit 114 as tabular data. For example, it is possible to output tabulated data as shown in FIGS. 10 and 12 in a tabular format.

  Next, the analysis graph creation unit 106 creates graph data of the summary data and analysis data stored in the summary data storage unit 113 and the analysis result storage unit 114. For example, FIG. 15 shows the ratio of the number of respondents by attribute stored in the aggregate data storage unit 113 as a pie chart.

  The analysis graph creation unit 106 creates a four-quadrant graph based on the importance value, the satisfaction value, and the attention value among the analysis data stored in the analysis result storage unit 114.

  For example, a four-quadrant graph is created using importance values and satisfaction values analyzed using the first numerical data stored in the analysis result storage unit 114. With the horizontal axis of the four-quadrant graph as the importance value and the vertical axis as the satisfaction value, a plot is made as to which position on the graph each total item corresponds to.

  For example, in FIG. 12, since the analysis item “noise” has an importance value of −0.20738 and a satisfaction value of 2.5, the analysis is performed at the position A in the graph shown in FIG. The item “Noise” is plotted. Similarly, other analysis items are plotted on a four-quadrant graph according to the importance value and the satisfaction value.

  In this way, by analyzing the importance value and the average value of each evaluation item on the 4 upper limit graph, it is analyzed which item of each evaluation item is supposedly important or satisfied. Can do. For example, evaluation items that are important but are not satisfied on average can be known at a glance.

  In FIG. 16, the size of the points plotted on the graph is plotted with the same shape and the same size for any evaluation item. Here, the size of the plot points may be changed according to the ratio of the attention value of each evaluation item.

  For example, FIG. 17 is a display example of a graph when the size of each evaluation item plotted is changed according to the attention value of the analysis data of FIG. 12 stored in the analysis result storage unit 114. In this way, by changing the size of the plot according to the attention level value, it is possible to analyze how much the evaluation item is relatively focused. For example, although the importance is relatively low, it is possible to know at a glance evaluation items with a high degree of attention.

  Further, FIG. 16 shows that a four-quadrant graph is created based on the first satisfaction value and the importance value, but the position based on the second satisfaction value and the importance value is displayed on the same graph. In the quadrant, the shape is displayed so that it can be understood how much the point position of the second satisfaction value and importance value is different from the point position of the first satisfaction value and importance value of the same evaluation item. You can also create a graph.

  For example, FIG. 18 shows a case where the change of the plot position is expressed by an arrow in order to understand the displacement from the point position of the first satisfaction value and importance value to the point position of the second satisfaction value and importance value. It is an example. By expressing the displacement with the arrows in this way, for example, when the arrows are directed in a direction of low satisfaction, this indicates that there is a risk that the satisfaction value of the evaluation item is lowered.

  As described above, the analysis result output unit 105 and the analysis graph 106 create an analysis table and an analysis graph, and display them on the questionnaire analysis report 600. FIG. 19 shows an example of a questionnaire analysis report 600. By confirming such an analysis table, the person who conducted the questionnaire can easily determine which evaluation items should be improved.

  This completes the explanation of each component of the questionnaire analysis system. Next, the process of the questionnaire analysis apparatus will be described in more detail using a flowchart.

  FIG. 20 shows the overall processing flow of the questionnaire analysis system. First, the questionnaire analysis system 100 accepts input of answer data at the answer data input accepting unit 101 (S2001). The answer data input accepting unit 101 accepts, as the questionnaire answer data 700, for example, the answer data converted into text data in CSV format as input data. The accepted answer data is stored in the answer data storage unit 111.

  After receiving the input of the answer data and storing the received answer data in the answer data storage unit 111, the weight correction unit 102 weights the answers of the answer data options and digitizes the options (S2002). ). The digitization of the options is performed in accordance with the weights of the answer options of the questionnaire stored in the questionnaire storage unit 110, and further according to the weights based on the keywords stored in the keyword storage unit 112. Change numeric weighting. The method for digitizing and weighting options by the weight correction unit 102 is as described above in the description of the weight correction unit 102.

  The weighting correction unit 102 calculates first numerical data and second numerical data. The first numerical data is numerical data obtained by digitizing the options corresponding to the weights of the answer options of the questionnaire stored in the questionnaire storage unit 110. The second numerical data is numerical data obtained by digitizing the first numerical data by further changing the weight based on the keyword recorded in the keyword storage unit 112. FIG. 21 is a processing flow in which the first numerical data is digitized by further changing the weighting based on the keywords recorded in the keyword storage unit 112 and calculating the second numerical data. The processing flow for calculating the second numerical data will be described later with reference to FIG.

  When the answer data options are weighted and the first numerical data and the second numerical data can be calculated, the answer data stored in the total data storage unit 113 is summed (S2003). The answer data totaling process is performed by the answer data totaling unit 104. The answer data totaling method by the answer data totaling unit 104 is as described above. The answer data is aggregated by summing up the entire answer data and by attribute attribute. Aggregation is performed for each question type corresponding to each question, which is the answer selection item that is the answer to the question of the identifier question. The answer selection items are tabulated for each option corresponding to each question stored in the questionnaire storage unit 110.

  Once the answer data has been aggregated, the importance value is calculated (S2004). The importance value is calculated by the importance value calculator 1031. The importance value calculation method by the importance value calculation unit 1031 is as described above. For the answer data stored in the aggregate data storage unit 113, the numerical value data of the answer data of the question whose question type is “overall evaluation” is used as the objective variable, and the answer data of the question data of other question types is used. Multiple regression analysis is performed by using the digitized data of the choice as a dependent variable, and a dependent variable for each question type is calculated. The dependent variable calculated here is used as the importance value of each question type. The calculation of the importance value is performed based on the first numerical data and the second numerical data stored in the total data storage unit 113, and the first importance value and the second importance value are calculated. Calculate the value.

  Once the importance value is calculated, the average value is calculated (S2005). The average value is calculated by the average value calculation unit 1032. The average value calculation processing by the average value calculation unit 1032 is as described above. The average value is calculated based on the first numerical data and the second numerical data for each question type. The average value may be calculated for each attribute. The higher the weighting value of the option is, the higher the satisfaction level is, the higher the satisfaction level, the higher the average value of the question type.

  Once the average value is calculated, the attention level value is calculated (S2006). The attention level value is calculated by the attention level value calculation unit 1033. The attention value calculation processing by the attention value calculation unit 1033 is as described above, but a more detailed attention value calculation processing flow will be described later based on the flow of FIG.

  If the importance value, the average value, and the attention value are calculated, an analysis result is output and an analysis graph is created (S2007). The analysis result is output by the analysis result output unit 105, and the analysis graph is generated by the analysis graph generation unit 106. The analysis result output processing in the analysis result output unit 105 and the analysis graph creation processing by the analysis graph creation unit 106 are as described above.

  For example, the analysis result output unit 105 creates tabular data such as the number of respondents by attribute based on the data stored in the total data storage unit 113. In addition, the analysis graph creation unit 106 is based on the total data and analysis data stored in the total data storage unit 113 and the analysis result storage unit 114. A 4-quadrant graph is created for the degree value and the attention degree value. The four-quadrant graph is, for example, a graph in which the position on the graph corresponds to each total item, with the horizontal axis representing the importance value and the vertical axis representing the satisfaction value. As described above, not only the four-quadrant graph is created based on the first satisfaction value and the importance value, but also the four-quadrant graph is created based on the second satisfaction value and the importance value. The satisfaction value and importance value change predictions may be represented on the graph by arrows or the like, or the size of the plotted points may be changed according to the attention value.

  When the analysis result output unit 105 and the analysis graph creation unit 106 create an analysis table and an analysis graph, the analysis table and the analysis graph are output on the questionnaire analysis report 600, and the questionnaire analysis report 600 is created (S2008). ). As described above, the questionnaire analysis report 600 may be displayed on a display monitor, or may be printed as a paper report as a form report. FIG. 19 shows an example of an analysis report output on a paper form as an analysis report.

  The above is the description of the overall process of the questionnaire analysis system. Next, among the processes performed by the questionnaire analysis system, first numerical data, which is numerical data that has been digitized in accordance with the weights of the answer options of the questionnaire stored in the questionnaire storage unit 110, is stored in the keyword storage unit. The process of changing the weight as the second numerical data by further changing the weight based on the keyword recorded in 112 will be described in detail.

  FIG. 21 is a processing flow for changing the weighting of the first numerical data to the second numerical data. In the process of correcting the weighting to the second numerical data, first, a list of answer data of each respondent whose weighting has been changed to the first numerical data is acquired from the total data storage unit 113 (S2101). If the list of answer data is acquired, the process is repeated in order for each answer data (S2102). Once the answer data to be processed is specified, a list of question types is acquired from the questionnaire storage unit 110. For example, a list of question types such as “noise”, “air environment”, “office furniture”,...

  If a list of question types can be acquired, the process is repeated for each question type (S2104). In the process for each question type, first, a list of keywords corresponding to the question type to be processed is acquired from the keyword storage unit 112 (S2105). For example, when the question type to be processed is “noise”, from the keyword storage unit 112, the question type is a keyword corresponding to “noise”, such as “noisy”, “cannot concentrate”, “speaking voice”, etc. The keyword is acquired, and it is checked whether or not the acquired keyword is included in the comment of the answer data to be processed (S2106).

  If a keyword is included in the comment of the answer data to be processed (S2106: YES), the second numerical value is obtained by correcting the value of the weighted numerical value data of the question type option to be processed based on the weight of the keyword. Calculate the data. For example, in the answer data on the first line in FIG. 5, “the desk is wide but the chair is squeaky and noisy” is recorded in the comment. Since the comment includes the “noisy” keyword, the numerical value of the answer data of the question corresponding to “noise” that is the question type to be processed is corrected. If the correction value of the keyword “noisy” is registered as “−0.3” as shown in FIG. 8, the question type is a question corresponding to the “noise” according to the correction value. The numerical value of the answer data for X is corrected from “4” to “3.7”.

  If the weighting value for option quantification is higher, the answer is more positive, and the lower the value, the more negative the answer is. A weighting numerical value is set as a correction value, and a weighting numerical value that adds a numerical value is set as a correction value for a positive keyword.

  When no keyword is included in the comment of the answer data to be processed (S2106: NO), or when the second numerical data calculation process is completed, the process is performed for all question types acquired in S2103. Is confirmed (S2108). If there is an unprocessed question type (S2108: NO), the process returns to S2104, and the process is performed with the next question type.

  If all question types have been processed (S2108: YES), it is checked whether all the answer data acquired in S2101 have been processed (S2109). If there is unprocessed answer data (S2109: NO), the process returns to S2102 and the next answer data is processed. If the processing has been performed for all the answer data (S2109: YES), the processing for changing the weighting to the second numerical data is terminated. Here, for the numerical data for which the calculation of the second numerical data has not been performed, the value of the first numerical data is applied as the second numerical data.

  The above is the description of the processing flow for changing the weighting of the first numerical data to the second numerical data. In such a process of changing the weighting, it is possible to calculate the answer data closer to the respondent's image by changing the keyword and the weighting value stored in the keyword storage unit in a timely manner.

  Next, details of the attention value calculation method performed by the attention value calculation unit 1033 will be described. FIG. 22 is a processing flowchart of attention level value calculation processing performed by the attention level value calculation unit 1033. The attention level value is calculated for each question type. In the attention value calculation process, first, a list of question types is acquired from the questionnaire storage unit 110 (S2201). For example, a list of question types such as “noise”, “air environment”, “office furniture”,...

  If the list of question types can be acquired, the process is repeated for each question type (S2202). In the process for each question type, first, a list of keywords corresponding to the question type to be processed is acquired from the keyword storage unit 112 (S2203). For example, when the question type to be processed is “air environment”, from the keyword storage unit 112, the keyword corresponding to the “air environment” is “hot” “cold” “temperature difference”.・ Keywords such as ・ are acquired.

  When the keyword list is acquired, comments of all answer data are acquired from the answer data storage unit 111 (S2204). Then, it is confirmed whether or not the keyword acquired in S2203 is included in the acquired comment, and the number of keywords included is counted (S2205). For example, when the answer data as shown in FIG. 4 is taken as an example, the keywords “hot”, “cold”, and “temperature difference” acquired with the question type “air environment” are included in the comment. Then, “3” is counted as the number of keywords.

  Then, the counted number of keywords is recorded in the analysis result storage unit 114 as an attention value corresponding to the question type (S2206). When the attention level value recording of the question type to be processed is completed, it is confirmed whether or not the processing has been performed for all the question types acquired in S2201 (S2207). If there is an unprocessed question type (S2207: NO), the process returns to S2202, and the next question type is processed. When all the question types have been processed (S2207: YES), the calculation of the attention level value is completed.

  The above is the details of the attention value calculation processing performed by the attention value calculator 1033. Thus, in the process of calculating the attention level value, the keyword stored in the keyword storage unit is registered mainly for the keyword that will be noticed at the time of conducting the questionnaire for each question type of the questionnaire. Thus, it is possible to effectively measure what the questionnaire respondents are paying attention to.

  In the above-described processing flow, the keyword storage unit 112 referred to by the weighting correction unit 102 and the attention level calculation unit 1033 has been described using the same data structure. However, the weighting correction unit 102 and the attention level calculation unit 1033 By recording the data in the keyword storage unit 112 to be referenced as separate data, it is possible to set keywords that are more suitable for the actual calculation of each numerical value.

  The configuration and processing of the questionnaire analysis system according to the present invention have been described above with reference to the drawings. However, each component of each illustrated device is functionally conceptual and is not necessarily physically illustrated. It doesn't need to be the same. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured.

  In the above embodiment, the explanation is based on the questionnaire analysis system. However, the above embodiment is a method for causing a computer that can execute various processes by hardware logic or a program to execute the process. The program may be such that a program prepared in advance for executing the above embodiment is read by a computer and executed.

DESCRIPTION OF SYMBOLS 100 Questionnaires analysis system 101 Answer data input reception part 102 Weight correction part 103 Answer data analysis part 1031 Importance value calculation part 1032 Average value calculation part 1033 Attention degree value calculation part 104 Answer data totaling part 105 Analysis result output part 106 Analysis graph preparation Unit 110 Questionnaire storage unit 111 Answer data storage unit 112 Keyword storage unit 113 Total data storage unit 114 Analysis result storage unit 200 Questionnaire execution server 300 Questionnaire input terminal 400 Questionnaire answer sheet (paper data)
500 Questionnaire answer sheet (electronic data)
600 Questionnaire analysis report 700 Questionnaire response data

Claims (5)

A questionnaire analysis system that aggregates the responses of multiple respondents to a questionnaire question item consisting of multiple questions and analyzes the evaluation of the respondents of the evaluation items corresponding to the question items,
An answer data input accepting unit that accepts answer data input of an option and a comment character string each of the respondents answered each of a plurality of question items of the questionnaire,
An answer data storage unit for storing an option and a comment character string input to the answer data input receiving unit;
A keyword storage unit storing a plurality of keyword information extracted from the comment character string;
A weighting correction unit that converts the options stored in the answer data storage unit into first numerical data according to weighting;
Any one or more of the questionnaire question items are set as comprehensive evaluation question items, the first numerical data corresponding to the comprehensive evaluation question item options are set as objective variables, and the question item options other than the comprehensive evaluation question items are selected. By performing multiple regression analysis that derives multiple regression equations using the corresponding first numerical data as explanatory variables, the dependent variable of each question item other than the comprehensive evaluation question item is calculated, and the calculated dependent variable is comprehensively evaluated. An importance value calculation unit that calculates the first importance value of the evaluation item corresponding to each question item other than the question item;
An average value calculation unit for calculating a first average value from first numerical data corresponding to options of each question item other than the comprehensive evaluation question item of the questionnaire;
Attention level value for calculating the attention level value of the evaluation item related to the keyword by totaling the number of keywords stored in the keyword storage unit included in the comment character string stored in the answer data storage unit A calculation unit;
A questionnaire analysis system comprising: an analysis result storage unit that stores the calculated first importance value, first average value, and attention level value for each evaluation item corresponding to a question item. The questionnaire analysis system according to claim 1,
In the case where the keyword stored in the keyword storage unit is included in the comment character string, the weight correction unit further includes the first numerical data of the options answered to the question item related to the keyword. Convert to second numerical data according to weighting,
The importance value calculation unit further uses the second numerical data corresponding to the options of the comprehensive evaluation question item as an objective variable, and the second numerical data corresponding to the options of each question item other than the comprehensive evaluation question item as an explanatory variable. Dependent variables of each question item other than the comprehensive evaluation question item are calculated by performing multiple regression analysis that derives the multiple regression equation as follows, and the calculated dependent variable corresponds to each question item other than the comprehensive evaluation question item As the second importance value of the evaluation item,
The average value calculation unit further calculates a second average value from second numerical data corresponding to options of each question item other than the comprehensive evaluation question item of the questionnaire,
The questionnaire analysis system characterized in that the analysis result storage unit is further an analysis result storage unit that stores the calculated second importance value and second average value for each evaluation item corresponding to the question item. The questionnaire analysis system according to claim 2,
The degree of attention of each evaluation item based on the first importance value and the first average value stored in the analysis result storage unit on the four-quadrant graph with the importance value and the average value as axes A questionnaire analysis system characterized by comprising an analysis graph creation unit that creates graph image data plotted with a size corresponding to the ratio of. The questionnaire analysis system according to claim 3,
The analysis graph creation unit further includes a first importance value based on a second importance value and a second average value stored in the analysis result storage unit for each evaluation item on the four-quadrant graph. And a graph analysis data plotted in a shape recognizing a change between each evaluation item and a plot of each evaluation item based on the first average value. A questionnaire analysis program for operating a computer as the questionnaire analysis system according to any one of claims 1 to 4.