Response data analysis device, response data analysis method, and response data analysis program

Abstract

The present invention reduces erroneous determinations when determining false answers. This answer data analysis device includes: a first calculation unit that calculates variability of answering times and variability of difficulties for a plurality of questions on the basis of the answering times and the difficulties for the respective questions; a second calculation unit that calculates a representative value of the answering time and a representative value of the difficulty for the plurality of questions on the basis of the answering times and the difficulties for the respective questions; and a determination unit that determines whether or not answers to the plurality of questions are false answers on the basis of an index value calculated by input of the variability of the answering times and the variability of the difficulties, and the representative value of the answering times and the representative value of the difficulties, to a function based on correlations with the false answers.

Description

[Technical field]

[0001] The present disclosure relates to a response data analysis device, a response data analysis method, and a response data analysis program. [Background technology]

[0002] Web surveys, in which a questionnaire (multiple questions) is distributed online and survey results (a set of responses to the multiple questions) are collected, generally have a high number of negative responses and require appropriate screening.

[0003] On the other hand, a bad answer has a characteristic that the answer time is shorter than that of a normal answer, and a method for determining whether an answer is bad based on the answer time has been proposed that utilizes this characteristic. With this method, survey results that are determined to be bad answers can be appropriately excluded from the tabulation results. [Prior art documents] [Non-patent literature]

[0004] [Non-Patent Document 1] Keiichi Hayakawa, Kazuo Takamine, Fumiyasu Yamada, "An Attempt to Select Inappropriate Answers by Focusing on the Consistency of Answers (Using SD Method Data)", Abstracts of the National Research Presentation Conference of the Japan Society for Management Information, 2007 Fall National Research Presentation Conference, Japan Society for Management Information, 2007 [Non-Patent Document 2] Shinya Masuda, Takayuki Sakagami, Masahiro Morii, "Poor Responses and Response Time in Web Surveys," Proceedings of the 81st Annual Meeting of the Japanese Psychological Association, Japanese Psychological Association, 2017 Summary of the Invention [Problem to be solved by the invention]

[0005] However, there are individual differences in response time, and each respondent has a different tendency for the length of response time. Therefore, in cases where there is a large individual difference in response time, if a bad response is judged based only on the length of response time, there is a possibility that a normal response will be erroneously judged as a bad response.

[0006] The present disclosure aims to provide a response data analysis device, a response data analysis method, and a response data analysis program that reduce erroneous determinations when determining poor responses. [Means for solving the problem]

[0007] According to one aspect of the present disclosure, a response data analysis device includes: a first calculation unit that calculates a variation in answer time and a variation in difficulty level for a plurality of questions based on the answer time and the difficulty level for each question; a second calculation unit that calculates a representative value of the answering time and the representative value of the difficulty level for a plurality of questions based on the answering time and the difficulty level for each question; The test system further includes a judgment unit that judges whether or not the answers to the plurality of questions are poor answers based on an index value calculated by inputting the variation in answer time and the variation in difficulty, and the representative value of the answer time and the representative value of the difficulty into a function corresponding to the correlation with poor answers. Effect of the Invention

[0008] According to the present disclosure, it is possible to reduce erroneous determinations when determining whether a response is bad. [Brief description of the drawings]

[0009] [Figure 1] FIG. 1 is a diagram illustrating an example of a system configuration of a questionnaire result collection system. [Diagram 2] FIG. 2 is a diagram illustrating an example of a hardware configuration of the response data analysis device. [Diagram 3] FIG. 3 is a diagram illustrating an example of a functional configuration of the analysis unit. [Figure 4] FIG. 4 is a first diagram showing a specific example of the process of each part of the analysis unit. [Diagram 5] FIG. 5 is a second diagram showing a specific example of the process of each part of the analysis unit. [Figure 6] FIG. 6 is a flowchart showing the flow of the response data analysis process. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] Hereinafter, each embodiment will be described with reference to the accompanying drawings. In this specification and the drawings, components having substantially the same functional configurations are denoted by the same reference numerals, and redundant description will be omitted.

[0011] [First embodiment] <System configuration of the survey result collection system> First, a description will be given of a system configuration of a survey result collection system including a response data analysis device according to the first embodiment. Fig. 1 is a diagram showing an example of the system configuration of the survey result collection system.

[0012] 1, the survey result collection system 100 includes a mobile terminal 110 carried by a respondent 140, and a response data analysis device 120 that analyzes response data. In the survey result collection system 100, the mobile terminal 110 and the response data analysis device 120 are connected to each other via a network 130 so as to be able to communicate with each other.

[0013] Mobile terminal 110 accesses a predetermined site via network 130 and displays a questionnaire provided by response data analysis device 120 on a display screen. Mobile terminal 110 also accepts answers entered by respondent 140 to the displayed questionnaire (multiple questions). Mobile terminal 110 also transmits a group of answers entered to the multiple questions to response data analysis device 120 as the survey results.

[0014] In Fig. 1, reference numerals 111 and 112 denote an example of a plurality of questions. As shown in Fig. 1, the questionnaire includes a plurality of questions (in the example of Fig. 1, question 1, question 2, ...). In the questionnaire, each question is represented by letters, and a respondent 140 inputs an answer to each question by selecting a radio button corresponding to a desired option from a plurality of options.

[0015] Note that the questionnaire shown in FIG. 1 shows only a portion of the questionnaire, and the questionnaire includes two or more questions. In addition, the questionnaire shown in FIG. 1 shows a case where each question has three options, but the number of options for each question is not limited to three. In addition, the questionnaire shown in FIG. 1 shows that answers are input via radio buttons, but the method of inputting answers is not limited to radio buttons.

[0016] A response data analysis program is installed in the response data analysis device 120 , and the response data analysis device 120 functions as a collection unit 121 and an analysis unit 122 by executing the program.

[0017] The collection unit 121 provides online questionnaires to the respondents 140, and generates response data based on the received questionnaire results.

[0018] Specifically, when the collection unit 121 is accessed by the mobile terminal 110, it provides a questionnaire to the mobile terminal 110. Furthermore, while providing the questionnaire, the collection unit 121 measures and records the answering time for each question. Furthermore, the collection unit 121 analyzes the content of each question and calculates the question difficulty. In this embodiment, the collection unit 121 calculates the question difficulty by counting the number of phrases in each question. Furthermore, the collection unit 121 receives a group of answers entered by the respondent 140 as the questionnaire result.

[0019] Furthermore, the collection unit 121 generates answer data including the answering time and question difficulty of each question, and the survey results (a group of answers to a plurality of questions).

[0020] The analysis unit 122 acquires the response data generated by the collection unit 121, and determines whether the survey results (a group of responses to a plurality of questions) are bad responses or not (whether the responses are bad responses or normal responses).

[0021] Specifically, the analysis unit 122 determines from the acquired response data: - An indicator showing the variability of answer times across all questions; - An index showing the variation in difficulty of all questions, - An indicator showing the average response time for all questions, - An index showing the representative level of difficulty for all questions, and uses each index to calculate the degree of bad answers in terms of variation and the degree of bad answers in terms of a representative value. The analysis unit 122 also calculates an overall degree of bad answers (an example of an index value of a bad answer) by weighting and adding the degrees of each bad answer. Furthermore, the analysis unit 122 determines whether the survey results (a group of answers to multiple questions) are bad answers (whether they are bad answers or normal answers) based on the calculated overall degree of bad answers.

[0022] Furthermore, the analysis unit 122 stores the survey results (a group of responses to a plurality of questions) in the data storage unit 123 in association with the judgment results (normal or incorrect responses). The survey results stored in the data storage unit 123 are screened based on the associated judgment results and then tallied. In this way, the response data analysis device 120 can obtain appropriate tallied results from the survey results.

[0023] <Hardware configuration of response data analysis device> Next, the hardware configuration of the response data analysis device 120 will be described. Fig. 2 is a diagram showing an example of the hardware configuration of the response data analysis device. As shown in Fig. 2, the response data analysis device 120 has a processor 201, a memory 202, an auxiliary storage device 203, an I / F (Interface) device 204, a communication device 205, and a drive device 206. The hardware components of the response data analysis device 120 are connected to each other via a bus 207.

[0024] The processor 201 has various arithmetic devices such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), etc. The processor 201 reads out various programs (for example, a response data analysis program) installed in the auxiliary storage device 203 onto the memory 202 and executes them.

[0025] The memory 202 has a main storage device such as a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The processor 201 and the memory 202 form a so-called computer, and the processor 201 executes various programs read onto the memory 202, whereby the computer realizes various functional units (e.g., the collection unit 121, the analysis unit 122).

[0026] The auxiliary storage device 203 stores various programs and various data used when the various programs are executed by the processor 201. For example, the data storage unit 123 is realized in the auxiliary storage device 203.

[0027] The I / F device 204 is a connection device that connects the input device 210 and the output device 220 with the response data analysis device 120. The I / F device 204 receives various instructions given to the response data analysis device 120 by an administrator (not shown) of the response data analysis device 120 via the input device 210. The I / F device 204 also outputs, via the output device 220, the results of processing executed by each functional unit of the response data analysis device 120 in response to the various instructions.

[0028] The communication unit 205 is a communication device for communicating with the mobile terminal 110 via the network 130 .

[0029] The drive device 206 is a device for setting the recording medium 230. The recording medium 230 here includes media that record information optically, electrically, or magnetically, such as a CD-ROM, a flexible disk, a magneto-optical disk, etc. The recording medium 230 may also include semiconductor memories that record information electrically, such as ROM, flash memory, etc.

[0030] The various programs to be installed in the auxiliary storage device 203 are installed, for example, by setting the distributed recording medium 230 in the drive device 206 and reading out the various programs recorded in the recording medium 230 by the drive device 206. Alternatively, the various programs to be installed in the auxiliary storage device 203 may be installed by being downloaded from the network 130 via the communication device 205.

[0031] <Details of the analysis section's functional configuration> Next, the details of the functional configuration of the analysis unit 122 will be described. Fig. 3 is a diagram showing an example of the functional configuration of the analysis unit. As shown in Fig. 3, the analysis unit 122 has a data acquisition unit 310, a variation index calculation unit 321, a representative value index calculation unit 322, a variation poor response degree calculation unit 331, a representative value poor response degree calculation unit 332, a determination unit 340, a threshold input unit 341, and a storage control unit 350.

[0032] The data acquiring unit 310 acquires the response data from the collecting unit 121. Furthermore, the data acquiring unit 310 notifies the variation index calculating unit 321, the representative value index calculating unit 322, and the storage control unit 350 of the acquired response data.

[0033] The variation index calculation unit 321 is an example of a first calculation unit, and calculates "an index indicating the variation in answer time for all questions" and "an index indicating the variation in question difficulty for all questions" from the notified answer data. Note that in this embodiment, a standard deviation is used as an index indicating the variation, but the index indicating the variation is not limited to the standard deviation and may be, for example, a variance value.

[0034] Furthermore, the variability index calculation unit 321 notifies the variability poor answer degree calculation unit 331 of the calculated "index indicating the variability in answering time for all questions" and "index indicating the variability in question difficulty for all questions".

[0035] The representative value index calculation unit 322 is an example of a second calculation unit, and calculates "an index indicating a representative value of answer times for all questions" and "an index indicating a representative value of question difficulty for all questions" from the notified answer data. Note that in this embodiment, an average value is used as an index indicating a representative value, but the index indicating a representative value is not limited to the average value and may be, for example, a median.

[0036] In addition, the representative value index calculation unit 322 notifies the representative value poor answer degree calculation unit 332 of the calculated “index indicating the representative value of the answer time for all questions” and “index indicating the representative value of the question difficulty for all questions.”

[0037] The variability poor answer degree calculation unit 331 calculates the degree of poor answering in terms of variability (variability poor answer degree) based on the notified "index indicating the variability in answering time for all questions" and "index indicating the variability in question difficulty for all questions."

[0038] In general, poor answers tend to have smaller variation in answering time, and there is a negative correlation between the degree of poor answers and the variation in answering time. In addition, the higher the degree of poor answers, the more negative the correlation between the variation in answering time and the variation in question difficulty, and the negative correlation between the variation in answering time and the degree of poor answers, so there is a positive correlation between the variation in question difficulty and the degree of poor answers.

[0039] Therefore, by inputting the "index indicating the variation in answering time for all questions" and the "index indicating the variation in question difficulty for all questions" into a predetermined function that takes the above correlation into account, it is possible to calculate the degree of variation in poor answering (first index value) according to the above correlation. Specifically, in this embodiment, the degree of variation in poor answering psd(x) of the respondent 140 (respondent name="respondent x") is calculated using the first function shown in the following formula (1). (Equation 1) psd(x) = θsd(x) / Tsd(x) In the above formula (1), θsd(x) is an index indicating the variation in the difficulty of all questions for the respondent 140 (respondent name = "respondent x"), and Tsd(x) is an index indicating the variation in the answering time for all questions for the respondent 140 (respondent name = "respondent x").

[0040] However, the first function shown in the above formula (1) is just one example, and the function for calculating the degree of variability in poor answering psd(x) may be any other function as long as it has a negative correlation with the variability in answering time and a positive correlation with the variability in question difficulty.

[0041] The representative value poor answer degree calculation unit 332 calculates the degree of poor answer in terms of the representative value (representative value poor answer degree) based on the notified "index indicating the representative value of answer time for all questions" and "index indicating the representative value of question difficulty for all questions."

[0042] In general, poor answers tend to require shorter answer times, and there is a negative correlation between the degree of poor answers and the representative answer time. Also, the higher the difficulty of the question, the longer the answer time tends to be, so there is a positive correlation between the representative question difficulty and the degree of poor answers.

[0043] Therefore, by inputting the "index indicating the representative value of answering time for all questions" and the "index indicating the representative value of question difficulty for all questions" into a predetermined function that takes the above correlation into account, it is possible to calculate the representative value bad answering degree (second index value) according to the above correlation. Specifically, in this embodiment, the representative value bad answering degree pm(x) of the respondent 140 (respondent name="respondent x") is calculated using the second function shown in the following formula (2). (Formula 2)pm(x)=θm(x) / Tm(x) In the above formula (2), θm(x) is an index indicating a representative value of the difficulty of all questions in the questionnaire answered by the respondent 140 (respondent name = "respondent x"), and Tm(x) is an index indicating a representative value of the answering time for all questions in the questionnaire answered by the respondent 140 (respondent name = "respondent x").

[0044] However, the second function shown in the above formula (2) is just one example, and the function for calculating the representative value poor answer degree pm(x) may be any other function as long as it has a negative correlation with the representative value of the answer time and a positive correlation with the representative value of the question difficulty.

[0045] The judgment unit 340 calculates the overall degree of bad answers (overall degree of bad answers) p(x) of the respondent 140 (respondent name = "respondent x") by inputting the "variation bad answer degree psd(x)" and the "representative value bad answer degree pm(x)" into the function shown in the following equation (3). (Formula 3) p(x)=α×psd(x)+β×pm(x) In the above formula (3), α and β are weighting coefficients. When the overall defective response degree is calculated using either the "variation defective response degree psd(x)" or the "representative defective response degree pm(x)", for example, (α, β) = (1, 0) or (0, 1) is set. When the overall defective response degree is calculated by equally adding the "variation defective response degree psd(x)" and the "representative defective response degree pm(x)", for example, (α, β) = (0.5, 0.5) is set.

[0046] The determination unit 340 compares the calculated overall defective response degree with a threshold input from the threshold input unit 341, and determines that the answer is defective if the calculated overall defective response degree is equal to or greater than the threshold. Also, the determination unit 340 determines that the answer is normal if the calculated overall defective response degree is less than the threshold. Furthermore, the determination unit 340 notifies the storage control unit 350 of the determination result (normal response, defective response).

[0047] The threshold input unit 341 receives an input of a threshold used when the determination unit 340 determines whether a response is a failure response or a normal response based on the overall failure response degree, and notifies the determination unit 340 of the threshold.

[0048] The storage control unit 350 associates the questionnaire results included in the response data acquired by the data acquisition unit 310 with the determination results made by the determination unit 340 and stores them in the data storage unit 123.

[0049] <Specific examples of processing in each part of the analysis unit> Next, a specific example of the process of each unit (here, data acquisition unit 310 to threshold input unit 341) of analysis unit 122 will be described.

[0050] (1) Specific example of processing by the data acquisition unit 310 to the representative value index calculation unit 322 First, specific examples of the processing of the data acquisition unit 310, the variation index calculation unit 321, and the representative value index calculation unit 322 among the units of the analysis unit 122 will be described. Fig. 4 is a first diagram showing a specific example of the processing of each unit of the analysis unit.

[0051] 4, reference numeral 400 indicates a specific example of answer data for the respondent 140 (respondent name="Respondent x") acquired by the data acquisition unit 310. As indicated by reference numeral 400, the answer data includes information items such as "question number," "survey result," "answer time," and "question difficulty."

[0052] A number that identifies each question is stored in "Question Number." The example in Figure 4 shows that the number of questions in the questionnaire is N.

[0053] In the "survey result", a group of answers to a plurality of questions input by the respondent 140 (respondent name="respondent x") is stored.

[0054] The "answer time" field stores the time taken by the answerer 140 (answerer name = "answerer x") to answer each question. In the example of FIG. 4, T 1 ~T N This shows that it took about 10 minutes to complete the process.

[0055] The "question difficulty" field stores the difficulty level of each question (as described above, in this embodiment, the number of phrases in each question). In the example of FIG. 4, the number of phrases in each question is D 1 ~D N This shows that.

[0056] In FIG. 4, the reference numeral 411 denotes a time T stored in the “answer time” field by the variability index calculation unit 321. 1 ~T N 13 shows the standard deviation Tsd(x) of the response time of respondent 140 (respondent name = "Respondent x") to questions 1 to N calculated using the above.

[0057] In addition, in FIG. 4, the reference numeral 412 indicates that the variability index calculation unit 321 calculates the D stored in the “question difficulty” field. 1 ~D N The figure shows how the standard deviation θsd(x) of the question difficulty of questions 1 to N was calculated using the above.

[0058] In FIG. 4, the reference numeral 421 denotes the T 1 ~T N 13 shows the average answer time Tm(x) of the respondent 140 (respondent name = "respondent x") to questions 1 to N calculated using the above.

[0059] In addition, in FIG. 4, the reference numeral 422 indicates the representative value index calculation unit 322 calculating the D stored in the "question difficulty level" 1 ~D NThe figure shows how the average question difficulty value θm(x) of questions 1 to N was calculated using the above.

[0060] (2) Specific Example of Processing by the Variation Defective Response Degree Calculation Unit 331 to the Threshold Input Unit 341 Next, a description will be given of specific examples of the processing of the variation defective response degree calculation unit 331, the representative value defective response degree calculation unit 332, the determination unit 340, and the threshold input unit 341 in the analysis unit 122. Fig. 5 is a second diagram showing a specific example of the processing of each unit in the analysis unit.

[0061] In Figure 5, symbol 511 shows how the variation poor answer degree calculation unit 331 calculates the variation poor answer degree of the respondent 140 (respondent name = "Respondent x") based on the standard deviation Tsd(x) of the answer time and the standard deviation θsd(x) of the question difficulty.

[0062] Also, in Figure 5, symbol 512 indicates the state in which the representative value poor answer degree calculation unit 332 calculates the representative value poor answer degree of the respondent 140 (respondent name = "Respondent x") based on the average answer time Tm(x) and the average question difficulty θm(x).

[0063] In FIG. 5, reference numeral 520 indicates a state in which the determining section 340 calculates the total defect response degree p(x) based on the variation defect response degree psd(x) and the representative value defect response degree pm(x).

[0064] Furthermore, in FIG. 5, reference numeral 530 indicates a threshold value input by the threshold value input unit 341, and reference numeral 540 indicates a judgment result judged by the judgment unit 340 based on the total defect response degree p(x) and the threshold value.

[0065] As shown in the example of FIG. 5, when the total defective response degree p(x) is equal to or greater than the threshold, the judgment unit 340 judges the response to be defective, and when the total defective response degree p(x) is less than the threshold, the judgment unit 340 judges the response to be normal.

[0066] <Response data analysis process flow> Next, a description will be given of the flow of response data analysis processing by response data analysis device 120. Fig. 6 is a flowchart showing the flow of response data analysis processing.

[0067] In step S601, the collection unit 121 provides a questionnaire (a plurality of questions) to the mobile terminal 110 in response to an access from the mobile terminal 110.

[0068] In step S602, the collection unit 121 measures and records the answering time for each question while the survey is being provided, and calculates the difficulty level for each question. The collection unit 121 also receives the survey results (a group of answers to multiple questions), and generates answer data by associating the answering time and the question difficulty level.

[0069] In step S603, the analysis unit 122 acquires the answer data, and calculates an "index indicating the variation in answer time for all questions" and an "index indicating the variation in question difficulty for all questions" based on the acquired answer data.

[0070] In step S604, the analysis unit 122 calculates "an index indicating a representative value of answering times for all questions" and "an index indicating a representative value of question difficulty for all questions" based on the acquired answer data.

[0071] In step S605, the analysis unit 122 calculates a degree of variation in poor answering based on the "index indicating the variation in answering time for all questions" and the "index indicating the variation in question difficulty for all questions." The analysis unit 122 also calculates a degree of representative poor answering based on the "index indicating the representative value of answering time for all questions" and the "index indicating the representative value of question difficulty for all questions."

[0072] In step S606, the analysis unit 122 calculates a total defective response degree by weighting and adding the variation defective response degree and the representative value defective response degree. The analysis unit 122 also compares the calculated total defective response degree with a threshold value to determine whether the questionnaire result is a defective response (whether the questionnaire result is a defective response or a normal response).

[0073] In step S607, the analysis unit 122 stores the determination result (bad answer / normal answer) in the data storage unit 123 in association with the questionnaire results (a group of answers to a plurality of questions).

[0074] <Summary> As is clear from the above description, the response data analysis device 120 according to the first embodiment: Based on the answering time and question difficulty for each question, an index showing the variation in answering time across multiple questions and an index showing the variation in question difficulty are calculated. Based on the answering time and question difficulty for each question, an index showing the representative answering time for multiple questions and an index showing the representative question difficulty are calculated. An index showing the variation in answering time for multiple questions, an index showing the variation in question difficulty, and an index showing the representative value of answering time for multiple questions and an index showing the representative value of question difficulty are input into the function ((Formula 1) to (Formula 2)) according to the correlation with poor answers. From this, the degree of variation poor answers and the degree of representative poor answers are calculated. Based on the calculated variation defective answer degree and representative value defective answer degree, a total defective answer degree is calculated. Also, based on the calculated total defective answer degree, it is determined whether the answers to the multiple questions are defective answers or not.

[0075] In this way, the response data analysis device 120 according to the first embodiment calculates the degree of poor answers by taking into account the variation in response time in addition to the representative value of the response time, and further the variation and representative value of question difficulty. As a result, according to the first embodiment, even if there is a large individual difference in response time and the tendency of response time to be long or short differs for each respondent, the degree of poor answers to multiple questions can be calculated with high accuracy.

[0076] As a result, according to the first embodiment, it is possible to reduce erroneous determinations when determining whether a response is bad.

[0077] [Second embodiment] In the above first embodiment, the question difficulty level is calculated by counting the number of phrases in each question, but the question difficulty level may be calculated based on factors other than the number of phrases or by adding factors other than the number of phrases. Also, in the above first embodiment, the question difficulty level is calculated while the questionnaire is being provided, but the question difficulty level may be set for each question in advance before the questionnaire is provided.

[0078] In the above first embodiment, the answerer 140 is assumed to answer all questions. However, the answerer 140 does not necessarily answer all questions. When an unanswered question is included, the analysis unit 122 may, for example, set the answer time of the question to zero and calculate an "index indicating the variation in answer time for all questions." Alternatively, an index indicating the variation in answer time may be calculated excluding the question.

[0079] In the above first embodiment, the collection unit 121 of the response data analysis device 120 has been described as measuring the response time for each question and calculating the question difficulty level during the provision of the questionnaire. However, the measurement of the response time and / or the calculation of the question difficulty level may be performed by the mobile terminal 110, and the collection unit 121 may be configured to receive the execution results together with the questionnaire results. In this case, the collection unit 121 may be configured to receive the execution results together with the questionnaire results, or may be configured to receive them separately.

[0080] The present invention is not limited to the configurations shown in the above embodiments, combinations with other elements, etc. These points can be changed without departing from the spirit of the present invention, and can be appropriately determined according to the application form. [Explanation of symbols]

[0081] 100: Survey result collection system 110: Mobile terminal 120: Response data analysis device 121: Collection Department 122:Analysis Department 310: Data Acquisition Department 321: Variation index calculation unit 322: Representative value index calculation unit 331: Variation and defective response degree calculation unit 332: Representative value defective response degree calculation unit 340: Judgment section 341: Threshold input section 350: Storage control unit

Claims

1. a first calculation unit that calculates a variation in answer time and a variation in difficulty level for a plurality of questions based on the answer time and the difficulty level for each question; a second calculation unit that calculates a representative value of the answering time and the representative value of the difficulty level for a plurality of questions based on the answering time and the difficulty level for each question; a determination unit that determines whether or not answers to the plurality of questions are bad answers based on index values ​​calculated by inputting the variation in answering time and the variation in difficulty level, the representative value of the answering time, and the representative value of the difficulty level into a function corresponding to a correlation with a bad answer; A response data analysis device having the above structure.

2. The response data analysis device of claim 1, wherein the function according to the correlation with the poor response includes a first function that has a negative correlation with the variation in the response time and a positive correlation with the variation in the difficulty level, and a second function that has a negative correlation with the representative value of the response time and a positive correlation with the representative value of the difficulty level.

3. 3. The response data analysis device of claim 2, wherein the index value is calculated by weighting and adding a first index value calculated by inputting the variation in response time and the variation in difficulty into the first function, and a second index value calculated by inputting a representative value of the response time and a representative value of the difficulty into the second function.

4. The response data analysis device according to claim 1 , wherein the level of difficulty is a number of phrases in each question.

5. The response data analysis device according to claim 1 , wherein the variation is a standard deviation or a variance, and the representative value is a mean value or a median value.

6. 2. The response data analysis device according to claim 1, wherein the responses to the plurality of questions are stored in association with the determination results by the determination unit.

7. a first calculation step of calculating a variation in answering time and a variation in difficulty level among a plurality of questions based on the answering time and the difficulty level of each question; a second calculation step of calculating a representative value of the answering time and the representative value of the difficulty level for the plurality of questions based on the answering time and the difficulty level for each question; a determination step of determining whether or not the answers to the plurality of questions are bad answers based on index values ​​calculated by inputting the variation in answering time and the variation in difficulty level, the representative value of the answering time, and the representative value of the difficulty level into a function corresponding to a correlation with a bad answer; A method for analyzing response data, the method being performed by a computer.

8. a first calculation step of calculating a variation in answering time and a variation in difficulty level among a plurality of questions based on the answering time and the difficulty level of each question; a second calculation step of calculating a representative value of the answering time and the representative value of the difficulty level for the plurality of questions based on the answering time and the difficulty level for each question; a determination step of determining whether or not the answers to the plurality of questions are bad answers based on index values ​​calculated by inputting the variation in answering time and the variation in difficulty level, the representative value of the answering time, and the representative value of the difficulty level into a function corresponding to a correlation with a bad answer; A response data analysis program for running the above on a computer.

Images