SURVEY QUESTIONS

CROSS-REFERENCE TO RELATED APPLICATION(S)

[0001] The applicants hereby claim the benefit of U.S. Provisional Patent Application No. 61 /552,912 filed on October 28, 201 1 , which is hereby incorporated by reference in its entirety.

BACKGROUND

[0002] In a survey, a group of questions is posed to each of a number of different survey participants, and the answers provided by the survey participants are noted. In some cases, the results of a survey are used merely to understand the views of the survey participants themselves, while in others, the results are attributed to one or more larger groups of which the survey participants are believed to be representative.

BRIEF DESCRIPTION OF THE DRAWINGS

[0003] Figure 1 is a high-level block diagram showing a typical environment in which a software, hardware, and/or firmware facility implementing the functionality described herein operates in some embodiments.

[0004] Figure 2 is a flow diagram showing steps typically performed by the facility in some embodiments in order to identify people likely to have responded accurately to survey questions.

[0005] Figure 3 is a flow diagram showing steps typically performed by the facility in some embodiments in order to construct a survey.

[0006] Figure 4 is a user interface diagram showing a sample display presented by the facility in some embodiments to pose fabrication gauge, unlikely, and likely survey questions.

[0007] Figure 5 is a flow diagram showing steps typically performed by the facility in order to discard survey responses for respondents least likely to have responded accurately to the survey. [0008] Figure 6 is a data structure diagram showing sample contents of a typical survey response table used by the facility in some embodiments to store survey responses.

DETAILED DESCRIPTION

[0009] The inventors have observed in conducting surveys that some survey participants respond inaccurately to survey questions, in some cases to qualify for some compensation provided in exchange for participating in the survey, or because responding to the survey inaccurately can be done in less time than responding to the survey accurately. The inventors have determined that such inaccurate responses tend to make survey results less useful and valuable. Accordingly, the inventors have developed a software and/or hardware facility for identifying people likely to respond accurately to survey questions ("the facility").

[0010] The facility poses a small set of qualification questions to an assembled group of prospective survey respondents. These qualification questions may either be posed to respondents as part of the survey that contains substantive survey questions, or as part of a "pre-survey" that a respondent completes before having an opportunity to participate in the main survey, or as part of a respondent router that assess honesty while also assessing which of a multiple of studies a respondent will qualify to complete. The set of qualification questions include a number of "unlikely" questions for each of which the correct answer is YES for a small percentage of the population, such as "Do you have more than 3 dogs as pets?" These "unlikely"— or low- incidence— questions should have low correlation, i.e., an accurate YES answer to one of the "unlikely" questions doesn't make more likely an accurate YES answer to any of the other "unlikely" questions. The set of qualification questions also includes one or more "fabrication gauge" questions for which a reliable percentage of the population is known for whom the correct answer is YES, such as "Were you born in the month of September?" The rate by which the YES answers to the "fabrication gauge" questions exceed the expected percentage provides an approximation of the portion of the group that is likely lying in an attempt to qualify for the survey, as participants may perceive that YES answers are more likely to result in their qualification for a survey. The facility discards people from the group in decreasing order of the number of "unlikely" questions to which they answered YES until the expected percentage of remaining undiscarded people answer YES to the fabrication gauge question. The people who remain in the group after this discarding are regarded by the facility as likely to respond accurately to survey questions. In some embodiments, the facility stores an indication that the remaining people are likely to respond accurately to survey questions. In some embodiments, the facility reports the survey responses of the remaining people in the group as the results of the survey come to the exclusion of the survey responses by people who do not remain in the group.

[0011] By behaving in some or all of these ways, the facility enables the operator of a survey to reduce the influence on survey results of inattentive, under-considered, malicious, or otherwise inaccurate survey responses.

[0012] Figure 1 is a high-level block diagram showing a typical environment in which a software, hardware, and/or firmware facility implementing the functionality described herein operates in some embodiments. The environment 100 includes a server computer system 150. The server computer system 150 includes a memory 160. The memory 160 includes software 161 incorporating both the facility 162 and data 163 typically used by facility. The memory further includes a web server computer program 166 for providing web pages and/or other information to other computers. While items 162 and 163 are stored in memory while being used, those skilled in the art will appreciate that these items, or portions of them, maybe be transferred between memory and a persistent storage device 173 for purposes of memory management, data integrity, and/or other purposes. The server computer system 150 further includes one or more central processing units (CPU) 171 for executing programs, such as programs 161 , 162, and 166, and a computer-readable medium drive 172 for reading information or installing programs such as the facility from tangible computer-readable storage media, such as a floppy disk, a CD-ROM, a DVD, a USB flash drive, and/or other tangible computer-readable storage media. The computer system 150 also includes one or more of the following: a network connection device 174 for connecting to a network (for example, the Internet 140), an information input device 175, and an information output device 176. In some embodiments, the facility operates on the server computer system to perform some or all of the following activities: designing a survey; administering a survey; identifying survey responses that are unlikely to have been prepared accurately; and/or compiling survey responses. [0013] The block diagram also illustrates several client computer systems, such as client computer systems 1 10, 120, and 130. Each of the client computer systems includes a web client computer program, such as web clients 1 1 1 , 120, and 131 , for receiving web pages and/or other information in response to requests to web server computer programs, such as web server computer program 166. The client computer systems are connected via the Internet 140 or a data transmission network of another type to the server computer system 150. Those skilled in the art will recognize that the client computer systems could be connected to the server computer system 150 by networks other than the Internet, however. In some embodiments, some or all of the client computer systems are used to complete a survey. In some embodiments, these client computer systems can include other server computer systems, desktop computer systems, laptop computer systems, mobile phones, personal digital assistants, tablet computers, televisions, cameras, automobile computers, electronic media players, etc.

[0014] While various embodiments are described in terms of the environment described above, those skilled in the art will appreciate that the facility may be implemented in a variety of other environments including a single, monolithic computer system, as well as various other combinations of computer systems or similar devices connected in various ways. In various embodiments, a variety of computing systems or other different client devices may be used in place of the web client computer systems, such as mobile phones, personal digital assistants, televisions, cameras, etc.

[0015] Figure 2 is a flow diagram showing steps typically performed by the facility in some embodiments in order to identify people likely to have responded accurately to survey questions. In step 201 , the facility constructs a survey. The facility's performance of step 201 is discussed in greater detail below in connection with Figure 3.

[0016] Figure 3 is a flow diagram showing steps typically performed by the facility in some embodiments in order to construct a survey. In step 301 , the facility creates an empty survey containing no questions. In some embodiments, rather than creating an empty survey in step 301 , the facility uses a template containing certain common elements, such as boilerplate questions, formatting or other design elements, disclaimers and/or privacy policies, etc. (not shown). In step 302, the facility incorporates into the created survey a number of unlikely questions. These are questions that would each be accurately answered YES by a small percentage of the population, and that have low correlation. As one example, in some embodiments, the facility includes the unlikely questions shown below in Table 1 , and/or similar questions.

I have a hunting license.

I have a subscription to the opera.

I have written a unicycle more than 100 yards in the last 12 months.

I have 3 or more dogs as pets.

I have attended a live jazz concert in the last 3 months.

I have operated a chain saw in the past 12 months.

Table 1

In some embodiments (not shown), to add diversity, one or more likely questions is included among the unlikely questions and the answers to these questions are ignored. An example likely question is "I have a driver's license."

[0017] In step 303, the facility incorporates into the survey a fabrication gauge question for which a reliable percentage of the population is known who will accurately answer YES. One example is "I was born in November," a fabrication gauge question to which 1 /12 of respondents would be expected to accurately answer YES. Any other month may be used, as may zodiac birth signs, Asian animal birth years, month of birth of ones mother or father, left- or right-handedness, eye color, etc. In some embodiments, the facility uses an expected percentage of affirmative answers for a fabrication gauge question that is matched to the population expected to take the survey. In various embodiments, the facility incorporates various numbers of fabrication gauge questions into the survey.

[0018] In step 304, the facility incorporates substantive questions into the survey that seek the information that is the subject of the survey. These questions may be intended to obtain information from the respondent that is useful in its own right, or may be used to "route" the respondent to one or more surveys for which the respondent is qualified. After step 304, these steps conclude.

[0019] Those skilled in the art will appreciate that the steps shown in Figure 3 and in each of the flow diagrams discussed below may be altered in a variety of ways. For example, the order of the steps may be rearranged; some steps may be performed in parallel; shown steps may be omitted, or other steps may be included; a shown step may be divided into sub-steps, or multiple shown steps may be combined into a single step, etc.

[0020] Returning to Figure 2, in steps 202-204, the facility loops through each of a number of respondents. In step 203, the facility administers the survey constructed in step 201 to the respondent to obtain a survey response for the respondent. In general, the survey response for a respondent contains or otherwise identifies the answers specified by the respondent in response to each question posed in the survey. In some embodiments, the survey may have survey flow logic that prevents all of the questions of the survey from being posed to a respondent, and thus the respondent's response does not identify an answer for that question. The survey may be administered to the respondent in series, in parallel, via the web or another client-server computer arrangement, using a monolithic computer system, or manually, such as by telephone, handwriting, scantron bubble forms, etc.

[0021] Figure 4 is a user interface diagram showing a sample display presented by the facility in some embodiments to pose fabrication gauge, unlikely, and likely survey questions. The display 400 includes a compound question asking which of eight statements are true for the respondent. Question 406 is a fabrication gauge question; questions 401 , 402, 404, 405, 407 and 408 are unlikely questions; and question 403 is a likely question. The respondent selects next to each question the YES or NO checkbox that contains the appropriate answer to the question. In some cases, respondents wishing not to be disqualified for the survey may check many or all of the YES boxes in an attempt to avoid disqualification. They may do so, for example, in order to obtain various kinds of compensation provided to respondents who are not disqualified from a survey; have a full opportunity to express their views and have them considered by decision-makers; etc.

[0022] Returning to Figure 2, in step 204, if additional respondents remain to take the survey, then the facility continues in step 202 to administer the survey to the next respondent.

[0023] In step 205, the facility discards survey responses for respondents that are least likely to have responded accurately to the survey. The facility's performance of step 205 is discussed in greater detail below in connection with Figure 5. If the lie detector is in a separate pre-qualification, those who failed the lie detector will not be routed to market research studies.

[0024] Figure 5 is a flow diagram showing steps typically performed by the facility in order to discard survey responses for respondents least likely to have responded accurately to the survey.

[0025] In step 501 , the facility creates an empty list of discarded responses and sets the value of index to the total number of unlikely questions included in the survey, which is six in the example. The facility continues to loop through steps 502-505 while the percentage of positive responses to the fabrication gauge question or questions among survey responses that have not been discarded in step 503 below exceeds the statistical expectations for positive responses to those fabrication gauge questions. In the example depicted above, there is one fabrication gauge question, asking whether it is true that "I was born in November." The facility attributes a 1 /12 or 8.33% statistical expectation for positive responses to these questions. Where four of sixteen or 25% of respondents answer affirmatively to this question, the percentage of positive responses to the fabrication gauge questions among undiscarded survey responses initially exceeds statistical expectations by 16.67%.

[0026] In step 503, the facility adds to the list of discarded responses any responses that respond positively to a number of unlikely questions that exactly matches the current value of the index. In step 504, the facility decrements the index. In step 505, if the percentage of positive responses to the fabrication gauge questions among undiscarded survey responses still exceeds the statistical expectations for positive responses to those fabrication gauge questions, then the facility continues in step 502 to perform another iteration of steps 502-505. After step 505, these steps conclude. A detailed example of how responses are discarded is discussed below in connection with Figure 6.

[0027] Figure 6 is a data structure diagram showing sample contents of a typical survey response table used by the facility in some embodiments to store survey responses. The table 600 is made up of a number of rows 601 -616 each corresponding to a person who completed the survey. Each row is divided into the following columns: a respondent id column 651 containing an identifier that identifies the respondent to which the row corresponds; unlikely question answer columns 652, 653, 655, 656, 658, and 659, each indicating the answer the respondent gave to an unlikely question; a dummy question column 654 indicating the respondent's answer to a likely or "dummy" question included in the survey; a fabrication gauge question answer column 657 indicating the answer given by the respondent to the fabrication gauge question; one or more substantive question answer columns 660 each indicating the respondent's answer to one of the substantive questions of the survey; and a total unlikely column 661 containing a count of the total number of affirmative answers given by the respondent to unlikely questions. For example, row 601 indicates that the respondent having respondent id 12990130 answered YES to unlikely question 1 a, YES to unlikely question 1 b, YES to dummy question 1 c, YES to unlikely question 1 d, YES to unlikely question 1 e, YES to fabrication gauge question 1 f, YES to unlikely question 1 g, YES to unlikely question 1 h, and 46 to substantive question 2, and that the total number of affirmative answers to unlikely questions given by this respondent is 6.

[0028] While Figure 6 shows a table whose contents and organization are designed to make it more comprehensible by a human reader, those skilled in the art will appreciate that actual data structure used by the facility to store this information may differ from the table shown, in that it, for example, may be organized in a different manner; may contain more or less information than shown; may be compressed and/or encrypted; may contain a much larger number of rows and/or columns than shown; etc.

[0029] In some embodiments, the facility discards responses in the example shown in Figure 6 as shown in Table 2 below.

As a result, the responses corresponding to rows 601 , 607, 609, and 610 are included in the discarded list. [0030] Returning to Figure 2, in step 206, the facility stores an indication that the survey responses not discarded is expected to be relatively accurate. In various embodiments, this involves storing an indication in the rows of the discarded responses that they have been discarded; storing an indication in the rows not discarded that they were not discarded; constructing a list of discarded responses; constructing a list of responses not discarded, etc.

[0031] In step 207, the facility reports the survey responses that were not discarded as the survey's result. In general, this involves reporting contents of substantive question columns of the table for rows not discarded, to the exclusion of the discarded rows. For example, if a substantive question 2 asks the respondent's age, then a survey result may be reported for this column that averages the answers to this question contained in non-discarded rows, that is, rows 602-606, 608, and 61 1 - 616, or 18, 35, 23, 25, 71 , 54, 31 , 29, 22, 34, 19, and 44 for an average of 33.75. Those skilled in the art will recognize that a variety of other approaches to reporting survey results in light of responses discarded may be employed.

[0032] The inventors have noted a variety of other factors that tend to predict that a person will respond inaccurately to a survey, such as a large number of surveys previously completed, large incentives for completing a survey, and cultural factors. In various embodiments, the facility combines different factors as part of selecting the survey responses to discard.

[0033] In addition, in some embodiments, the facility performs post lie detection validation testing using the methods described in U.S. Patent Application No. 12/764,838 filed on April 21 , 2010, which is hereby incorporated by reference in its entirety. In some embodiments, the facility uses the described techniques to compare two multi-dimensional data sets -- one with those who failed the honesty test and one with those who passed it -- to create post hoc a "lie detector." In this approach, the facility uses previously claimed questions from "The Standard" along with a couple of low incidence questions. People who fail these questions are generally outliers ("outliars") from the entire data set. The facility determines that their answers to the standard calibration questions don't fit the expected structure of the data set as a whole. The facility does this by calculating a forest on the calibration study (by creating a series of trees that predict each variable from every other variable, and removing the lie detector variables to avoid any problems of independence) and then calculating the proximity matrix (the number of times each pair of panelists is classified in the same node) and matching low proximity panelists, i.e. the panelists that don't look like any other panelists, with the number of low incidence questions that they answered positively.

[0034] In some embodiments, the facility uses high-incidence questions in place of some or all of the above-described low-incidence questions, and treats a NO answer to these as an aberrant answer, as contrasted with treating YES answers to low-incidence questions as aberrant answers. For example, in some embodiments, the facility uses the following high-incidence question, to which a NO answer is treated as aberrant: "Please answer yes or no: I have not purchased a minivan within the past two weeks."

[0035] It will be appreciated by those skilled in the art that the above-described facility may be straightforwardly adapted or extended in various ways. While the foregoing description makes reference to particular embodiments, the scope of the invention is defined solely by the claims that follow and the elements recited therein.