DOI: 10.31038/PSYJ.2023551

Abstract

Through the use of Mind Genomics coupled with AI (artificial intelligence), the researchers explored responses to an almost totally new topic, retail analytics, specifically what might be the applications in the business environment. AI generated four questions about the use of retail analytics, and subsequently four answers (elements) to each question. These AI-generated answers were combined into unique sets of 24 vignettes, one set for each of 50 respondents. The ratings of the respondents in terms of ‘interests me’ were then deconstructed into the contribution of each element to the rating. The deconstruction by regression analysis was done for total panel, for two viewpoints about what a company would do with the data (better prices; better service), and three mind-sets or ways of thinking about the topic (MS1 – Focus on price/profitability; MS2 – Focus on innovation/operations; MS3 – Focus on outside world behavior, MS 3). The paper concludes with AI analysis and summarization of the strong performing elements for each identified group, using the same six queries for summarization. The paper demonstrates the potential for rapid, insightful learning for new topics, a learning promoted by AI and by human ‘validation’ of the ideas generated.

Introduction

This paper presents an exploration into what might be considered a specialized topic, subjective responses to retail analytics.

“Retail analytics involves using software to collect and analyze data from physical, online, and catalog outlets to provide retailers with insights into customer behavior and shopping trends. It can also be used to inform and improve decisions about pricing, inventory, marketing, merchandising, and store operations by applying predictive algorithms.” [1]

As this definition shows, retail analytics involves extensive data collection, complex analysis and presentation so that executives can make decisions for the benefit of their retail operations. For the analytics, there is the analyzed. Today’s marketing savvy consumers know that data is collected from them, and they expect the data to be used for their benefit, viz. personalized services and an improved shopping experience [2]. It’s an implicit contract.

For years now, retailers and most brands, promote their goal to be “consumer-centric,” i.e., putting the consumer at the center of everything a retailer or brand does, in effect placing the consumer in the driver’s seat, a position allowing them to tell brands how to shape brand experiences for them. Oddly, when it comes to retail analytics, the conversations are about executives interpreting consumer data – sales, trends, satisfaction levels to name three, but we could not find examples where the views of consumers about the use of – and how to use – retail analytics, which is derived from their data, to improve their shopping experiences. This gap provided the novel opportunity to contribute new knowledge by turning the research of subjective responses to retail analytics on its head – to explicate the bottom-up perspectives of consumers to executives, leaders who may then become even more consumer-centric in ways that benefit their operations and improve their customers’ experiences.

Our modern age continues to grow in technological capabilities, empowered as it is by the computer, by the Internet, by the so-called Internet of Things, and by the hyperfocus on optimizing in real time. One can scarcely go through a day without being exposed to ceaseless streams of advertisements and calls to take a buying action. Often, the advertisements are for items already purchased, for items recently viewed or for items abandoned in an e-commerce shopping cart, such advertisements and calls to act emerging from rapid, micro-second analysis of people’s shopping behaviors. Indeed, the analytic capabilities are so strong that the previous fad for gaining substantial insight, big data, looks almost antiquated in terms of the ability to deal with small, immediate, personalized data, generated nonstop.

It is no wonder that the world is awash in data. Our ability to formulate scientific questions, to track trends, to subject this rapid life to the slow majesty of scientific inquiry seems to vanish as the speeds and volumes of data feeds increase to an accelerating beat. It is difficult, indeed, to, ‘think slowly’ while in in the throes of massive data, massive opportunities to optimize.

We are accustomed to the slow, majestic, ingrained, now entrenched system of hypothetico-deductive reasoning [3]. The basic idea is that science, or perhaps even people personally, ‘advance’ by forming a hypothesis about something, and rigorously testing that hypothesis, attempting to falsify Whatever is not falsified begins to have the ring of truth to it. At the same time as technology is speeding up data production and acquisition, there is a need to speed up knowledge acquisition and thinking. It may require exceptional innovation to create knowledge in the hard sciences, such as biology and chemistry, but to create valuable knowledge in the softer human-centered sciences may be less of a problem. Indeed, with the advances already made in computers some of the paradigm changes may be within today’s grasp.

Mind Genomics, the Promise of AI, and a Vision of the Future

The study is part of the new effort in Mind Genomics to accelerate the acquisition of knowledge and insight for the world of the everyday, and for topics involving human feelings about situations and activities that we often overlook because of their sheer invisibility, such as our present topic of subjective feelings towards retail analytics. There is another motivation as well, the desire to demonstrate through an ongoing research program, whether the combination of artificial intelligence with systematized testing among humans can reveal new aspects of daily life, and even point to weak signals of attitudes which are evolving.

The underlying science, Mind Genomics, is an emerging discipline with rooms in a number of areas, including experimental psychology to search for causation of behavior, statistics which allow researchers to work with combinations of independent variables in the way which often occurs in nature, and finally consumer research which looks at how people make decisions about the world of the ‘ordinary’, the quotidian reality in which people actually live, function, thrive or fail.

The history of how Mind Genomics emerged or better ‘evolved’ has been told before, in detail. The story is simple, involving the basic question of how people respond when asked to evaluate a compound stimulus comprising a variety of features. The question is simple, and even the thinking is rudimentary. The sole focus is simply to see ‘what happens.’ The effort is simply to find patterns in nature. There is no effort to create a theory, to prove or disprove a theory, although those noble efforts can certainly occur. The worldview underneath searching for patterns in nature come from psychophysics, the branch of experimental psychology, which searches for regularities in nature, such as the perceived sweetness of different concentration of sucrose in a water solution [4], or perhaps more of interest to industrial concerns, the perceived sweetness of different combinations and concentrations of artificial, high-potency sweeteners, in beverages. The effort is to discover relations in nature, regularities. The mass of such discoveries, especially in a defined and coherent field, becomes technological know-how and even the basis of science.

The project itself was run on an accelerated timetable, the total involvement being less than three hours, although spread over two periods, the first being the set up and empirical experiment, the second being the automated analysis using artificial intelligence. These two together constituted the period of less than three hours. The writing of the paper took considerably longer, but plans are underway to make the writing as quick as the design and the field research. In that way the Mind Genomics approach can be a living instantiation of what studies are, namely decision making in real time in the real world of the everyday. It is worth noting that the first use of AI in Mind Genomics goes back four years, with the process being long because it was a handcrafted combination of AI and Mind Genomics [5].

Part 1 – The Creation of the Study, and the Execution with Respondents

Mind Genomics studies are now scripted, following a templated procedure which reduces both the effort to acquire data as well as the ‘angst’ involved in doing an experiment. A half-century of experience by author Moskowitz in the world of science has continued to reveal that anxiety reported by individuals who are not scientists in their daily lives but are asked to ‘do science’ for a specific objective.

The scripting for Mind Genomics is set up to ensure that the researcher can provide the necessary information, in the proper format. The actual experiment involves combining relevant ‘messages’ about a topic (called ‘elements’) into small, easy to read ‘vignettes’, a vignette comprising 2-4 such messages. The combinations are then evaluated by people, using a rating scale. The respondents each end up evaluating 24 different vignettes, the vignettes for one respondent different from the vignettes for another respondent, but the set of elements being the same. Each respondent rates a unique set of 24 vignettes [6].

Mind Genomics studies are now scripted so that the studies can be set up on either a computer or a smartphone, run, and results downloaded with a short period of time. Table 1 provides key information about the study, information that will be explicated. The text in Table 1 comes directly from the report, contained in an Excel workbook, which can be downloaded either immediately for partial analysis (without the artificial intelligence summarizer), or after 30 minutes for complete analysis (with the artificial intelligence summarizer).

Table 1: Key information about the study provided by the Excel report.

The set-up begins with the naming of the study, and then the instruction to provide four questions germane to the study topic. The researcher is prompted to structure the four questions so that they ‘tell a story.’ During more than a decade, from the time that Mind Genomics was opened to the public as a ‘software as a service,’ there has been increasing number of situations where researchers seemed to have been overwhelmed by the task of creating elements. With the advent of artificial intelligence, such as Chat GPT, it has become easy to use AI to create questions. All the researcher has to do is write down a short paragraph in the Idea Coach ‘box’ in the Mind Genomics template, and the AI returns with 30 different questions.

Figure 1 shows the request for four questions (left panel), and the four questions provided by the researcher (right panel). Table 2 shows one run of the Idea Coach, returning with the 30 questions after the AI has been prompted by a short description of the project. The actual time for this first step with the Idea Coach is approximately one 60-90 seconds after the researcher provides the Idea Coach with the small description. The researcher can select 0-4 questions. The questions selected are automatically put into the template. The researcher can then repeat the action to get other answers. The actions are independent of each other, so that each use of the Idea Coach to provide questions is independent of the previous use. In this way a person new to the topic can learn a great deal about the topic, even without selecting questions. It is the sheer number of different questions presented in a short period of time which constitutes an ‘education’ for the researcher on the very topic being studied. It is for that reason that the approach is considered ‘Socratic.’

Figure 1: The templated request for four questions (left panel), and the four questions provided by the Idea Coach (right panel).

Table 2: The 30 questions returned by Idea Coach after being prompted by a small paragraph about the purpose of the study.

The next step consists of providing four answers to each question. Once again the Idea Coach comes into play. This time the information provided to Idea Coach is in the form of a query, that query created directly from the text of the specific question. When the researcher wants to change the query, it becomes a simple task of editing the question.

Figure 2 shows the formatted screen for Idea Coach, with the left panel showing the request for four answers, and the panel showing the completed set of answers used in the study. Table 3, in turn, shows 15 answers generated from one run of the Idea Coach, with the first question. The time elapsed in about 30 seconds.

Figure 2: The templated request for four answers for question #1 (left panel), and the four questions provided by the Idea Coach (right panel).

Table 3: The 15 answers returned by Idea Coach after being prompted by Question #1.

Once the study has been launched with the appropriate respondents, done easily within the BimiLeap program, the researcher selects the source of respondents, using the screen shown in Figure 3.

Figure 3: Source of respondents, selected by the researcher as the last part of the project set-up.

The respondent is ‘sourced’ from a panel provider specializing in on-line surveys. Across the world there are many such panel providers. These providers have lists of respondents and qualifications, individuals who have agreed to participate in similar studies for a reward provided by the supplier. The researcher need not know the agreement. All that is necessary is for the panel provider to source the correct respondent.

The Mind Genomics study can have elements (questions and answers) from different languages, and different alphabets, although the instructions for the actual set-up as done by the research are currently available in just a few languages (e.g., English, Chinese, Arabic Hebrew.

The actual experiment with the participant lasted approximately 3-5 minutes. The experiment begins with a short orientation provided by the researcher, obtains responses to the self-profiling questions (including age and gender, not shown here), and then presents the respondent with the 24 different, systematically designed vignettes. The vignettes comprise 2-4 answers (elements), at most one answer from a question, but often no answers from a question.

The experimental design ensures that the ratings from each respondent can be analyzed separately, as well as being able to be analyzed as part of the group. The vignettes are set up so that each individual evaluates a unique set of 24 vignettes, a design structure which enables the researcher to explore many different aspects of an issue, without having to choose what combination of elements give the best chance of discovery.

It is important at this juncture to keep in mind that the effort is more in the world of ‘hypothesis-generating’ than in the world of ‘hypothesis-testing.’ Quite often researchers really have no hypotheses to test but are constrained to do the study as if it were guided by a hypothesis. Mind Genomics disposes with that, freely representing that it is a discovery tool, to identify patterns which may be of interest in the way people think about a topic. The respondent sees an introduction to the study, usually presented in shortened form so that the specific information must come from the elements selected by the researcher. Figure 4 (top panel) shows the orientation question. Figure 4 (bottom panel) shows the five-point scale used by the researcher. Some of the text is cut off from the scale description. Figure 5 (top panel) shows the additional self-profiling questions as these appear to the respondent at the start of the evaluation by the respondent. Figure 5 (bottom panel) shows a clear reproduction of a vignette, as prescribed by an underlying experimental design. To the respondent there is no ’underlying pattern’ to be discovered. Rather, the vignette itself looks like it was randomly assembled, in virtually a haphazard manner.

Figure 4: Respondent orientation screen in set-up program (top panel), and rating scale specifics in set up program (bottom panel).

Figure 5: Screen shot for the pull-down menu of self-describing questions (top pane), and example of a four-element vignette plus rating question/scale as the respondent would see it on a full screen (bottom panel).

Database Structure, Analysis, and Reports – Total Panel

Almost fifty years ago, it was possible to accelerate the acquisition of data and even the preliminary analysis of such data. One could gather data quickly, and using mark sense cards and later electronic input one could analyze the data by programs then available, usually programs which specialized in descriptive analysis. The objective for this exercise is to move dramatically beyond that simple original analysis, pushing the limits by making use of smart experimental design, clustering, and applied artificial intelligence using GPT.

The actual data analysis is straightforward, made so by the judicious use of the experimental design, pre-selected so that the different sets of 24 combinations are really isomorphs of each other. Although the actual combinations, the vignettes, all differ from each other, mathematically the structure of each set of 24 vignettes is the same. In this way the researcher is assured of having a powerful analytic tool, which explores a great deal of the ‘design space’ (the possible combinations). Clearly, more individuals, more respondents means more of the design space is covered.

The analysis is made possible by the creation of a simple to use database. Each row of the database corresponds to one of the 24 vignettes tested by a respondent. Thus, for this study, comprising as it does the responses of 50 individuals, the database contains 50×24 or 1200 rows of data. The columns themselves are divided not bookkeeping columns (row number, study name, respondent number, test order number for the respondent, via 1-24), then 16 columns devoted to showing absence of element (value 0) or presence of element (value 1), and finally, the rating assigned, and the response time. Response time is defined as the number of seconds from the appearance of the vignette to the actual response assigned.

After the data are collected, the program creates new variables, specifically a binary variable TOP (ratings 5 and 4 transformed to 100; ratings 3, 2 and 1 transformed to 0), and a second binary variable BOT (ratings 1 and 2 transformed to 100, ratings 3, 4 and 5 transformed to 0). The BimiLeap program adds a vanishingly small random number (<10-4) to each BOT and TOP value, in order to ensure variation in the values of TOP and BOT. This prophylactic action ensures that there will be the necessary variability even when the respondent rated all vignettes either 1 or 2, or 3, 4 or 5,.

The first analysis generates an equation relating the presence/absence of the 16 elements to the newly created dependent variable TOP. Table 4 shows the parameters of the equation, expressed as: TOP = k0 + k1(A1) + k2(A2) … k16(D4). . The use of an experimental design, permuted across the 50 respondents to create 1200 vignettes, ensures that the OLS regression will not encounter any problems in terms of variables correlated with each other. The coefficients have absolute value, viz., a 5 is half the value of a 10. This is important. Often it is only differences in coefficient values which are relevant, e.g., when the experimental design calls for each vignette to have exactly one element from each question. That ‘tempting’ requirement substantially weakens the results.

Table 4: Elements for the Total Panel which drive TOP (Sounds interesting).

With the recognition of these properties of the regression results, the researcher can quickly understand the dynamics revealed in the experiment. The results are immediate and obvious once the meaning of the coefficients is understood. The coefficients tell us the proportion of times a response to the vignette will be 5 or 4 (here … ‘sounds interesting’) when the element is put into the vignette.

Note: Other binary dependent variables can be created

BOT = rating 1,2 transformed to 100. ‘Does not sound interesting’

RATE52 = rating 5,2 transformed to 100. ‘Could deliver’

RATE 41 = rating 4, 1 transformed to 100. ‘Won’t deliver’

Table 4 shows us 17 numbers, the additive constant and 16 coefficients. Mind Genomics studies generate a large number of coefficients. We are not interested in the coefficients which are 0 or lower. These low coefficients say that the presence of the element in the vignette ‘does not add’. It does not mean that the element actually detracts, or perhaps just as likely, the element is irrelevant, leading to a rating of 3.

The additive constant tells us the proclivity of the respondents to say ‘sounds interesting’ in the absence of any elements in the vignette. By underlying design, the vignettes all comprise a minimum of two elements and a maximum of four elements. Therefore, the additive constant is simply a correction factor in statistics. ON the other hand, we can use it as a baseline, a proclivity for the respondent to say, ‘sounds interesting’. We will use it to gain that insight. Table 4 shows the additive constant to be 70, meaning that when it comes to knowing that the topic will be store analytics, 70% of the respondents will be strongly positive. Had we done this type of experiment across years ago we could have measured the change in this additive constant to get a sense of how new ideas are accepted.

Only three of the 16 elements generate coefficients of 1 or above. The remaining generate coefficients of 0 or negative. For the sake of clarity, and for the sake of allowing patterns to come through, we do not show any 0 or negative coefficients. If we discard the element with coefficient 1, quite close to 0, we are left with the finding that of our 16 best guesses using artificial intelligence, edited by a human researcher, we end up with only two strong performing elements, and indeed only modest performers at that.

Shown these results and presented with the Mind Genomics approach for the first time, the critic of artificial intelligence might aver, perhaps even strongly aver, that these results contradict the claims of AI proponents that AI can be as good or better than people. The reality is not as positive, however. When a person selects the elements alone, quite often the person does about as poorly, or perhaps slightly better.

C4: Analytics and modeling can help improve transportation management to reduce shipping costs and lead times.

D3: Data analytics can be used to monitor social media activity, helping retailers identify trending products and capitalize on buzz.

Before moving on to an analysis of subgroups and mind-sets, it is instructive to see what AI can provide in terms of a standardized interpretation of the results. The analysis by AI should be considered simply as tentative observations made by a heuristic. It will still take a person to go through the data, but in the interests of speed, one might employ the AI heuristic to get a sense of the answers before spending time with the results.

Table 5 shows the response of AI to six queries. These queries are listed below.

1. Interested in
2. Create a label for this segment
3. Describe this segment
4. Describe the attractiveness of this segment as a target audience:
5. Explain why this segment might not be attractive as a target audience:
6. Which messages will interest this segment?

Table 5: AI first scan and interpretation of the strong performing elements (6 and higher) for the Total Panel, on six queries.

The queries look only at the moderate and higher elements, viz., those with coefficients of +5 or higher. Elements with coefficients of 4 or lower are ignored. For our data in Table 4 using the Total Panel the AI uses only two of the 16 elements to write its analysis.

Results from Self-profiling Questionnaire – What will the Store Do with the New Analytics

Our second analysis parallels the first, this time focusing on how the respondent feels about what the store will do with the analytic information. The actual question and answers appear below. Out of four possible answers, most of the respondents (43 of 50) chose only two, better price and better service, respectively.

If a store knows a lot about customers, do you think the store will

1=Offer better prices

2=Satisfy the customer more

3=Help make a better world

4=Use the information to its own advantage

Two groups or segments emerge, those who feel that the store will offer better prices using the analytics data (15 of 50), and those who feel that the store with use the information to better satisfy customers (29 of 50). Both have similar additive constants (63 and 69) but respondent in radically different ways to the elements.

The data for this new analysis comparing different points of view on what will the store do with the analytics appear in Table 6 for the coefficients, and Table 7for the AI summarization. Once again all coefficients of 1 or lower are not shown. In the interests of simplicity, all elements without at least one coefficient of ‘2’ Table 6 suggests that the response of those in Segment 1 (believe the store will offer better prices) are strong and focused, whereas there are no strong performing elements for Segment 2 (believe that the operations will be better) (Table 7).

Table 6: Elements which drive TOP (Sounds interesting) for the two segments emerging from the question self-profiling question: If a store knows a lot about customers, do you think the store will…

Table 7: First scan and interpretation of the strong performing elements (6 and higher) by for the two segments, based on the self-profiling question of what the store will do with the results from the data analytic.

Results from Dividing Respondents into Mind-Sets Based Upon the Pattern of Coefficients

Our final data analysis will focus on the creation of mind-sets, groups of respondents who are put together by the k-means clustering program [7], based upon the similarity of the patterns made by their coefficients. For the clustering we use all 16 coefficients, whether positive or negative, although we only show the positive coefficients in the results. Clustering does not use the additive constant.

The clustering program is embedded in the BimiLeap program, generating at first two mind-sets, and then totally once again, three mind-sets. Each respondent is assigned to only one of the two or three mind-sets. Furthermore, the mind-sets encompass all respondents.

The objective of the clustering exercise is to discover hitherto unexpected groups of respondents based upon the patterns of their coefficients from thousands of small studies clustering based on coefficients emerges with meaningful, interpretable groups of respondents, even though the process is purely mechanical and mathematical.

Table 8 shows the positive performing elements for the three-mind-solution. The two-mind-solution is not shown in the interest of space. Table 9 once again shows the AI interpretation of the results, obtained by applying six queries to each of the three mind-sets.

Table 8: Elements which drive TOP (Sounds interesting) for the three mind-sets, emerging from k-means clustering of all the element coefficients from the 50 respondents. Only those elements with at least one coefficient of 2 or higher are shown.

Table 9: AI interpretation of the strong performing elements (6 and higher) by AI for the three-mind-set solution.

In contrast to the data from the Total Panel, and the data from the Self-Defined Segments regarding what the company will do with the segments, the division by the pattern of coefficients shows strong performing coefficients, and meaningful stories which repeat from the two-mind-set solution (data not shown in the interest of space.).

Beyond Interpretation to Understanding Performance at a Glance

A continuing issue emerging from these small-scale studies is ‘How did we do?’ The same question is asked by researchers in most areas, with the underlying issue touching on ‘did the experiment work?’ With Mind Genomics coupled with AI at the very early stages, and available world-wide at the press of a key, the issue becomes the degree to which the study generated anything of value. The notion of ‘value’ is not the personal value of the data to the researcher, nor value in terms of scientific reproducibility and validity, but rather did the effort lead to any strong performing elements. We learn that when we have strong performing elements there is a strong link between an element and the rating question. That linkage is what the researcher seeks in these studies. Knowing that some elements get high scores and other elements get low scores tells the researcher she or he can move in the direction of the high scoring elements. It is in those elements that the relevant issue can be further understood, at least in the minds of the people who are the respondents.

As part of the effort to ‘systematize’ the use of Mind Genomics in this era of easy-to-use techniques empowered by AI, we present the IDT, Index of Divergent Thought.

The objective of IDT is to measure the impact of the elements. The IDT generates a simple index, shown in Table 10. Each positive element is weighted by the relative number of respondents when we consider the entire study as six groups (Total, Mind-Sets 1 and 2 in the 2-Mind-Set solution, and Mind-Sets 1, 2 and 3 in the 3-Mind-Set Solution). Each of the six groups has a specific number of respondents, which add up to 3x the base size, or 3×50, viz. 150. The IDT is simply the weighted sum of the positive coefficients, the weight being the ratio of the number of respondents in a group versus the total number of 150. Note that the two-mind-set solution was developed, but not shown in the interest of space.

Table 10: The IDT, Index of Diverged Thought, showing the performance of the elements, and thus the strength of the thinking behind the specific Mind Genomics study.

The ideal use of the IDT at these early days of implementation is to determine how good was the choice of elements. The IDT for this study is 33.05, on the low side. Keeping in mind that the elements were obtained from AI, and only modestly edited, we have with the IDT a way to quantify the strength of the ideas as they are perceived by people.

Discussion and Conclusions

As emphasized above, this study was undertaken as a study of an experience, the experience being the efforts to learn about a topic new to the senior author, viz., sustainability in the world of retail analytics. The original idea for this study came from a call for papers from the newly founded journal ‘Sustainability.’ The fact that the senior author had little experience with the topic of retail analytics, and indeed scarcely, if ever, thought about the topic, made the study more interesting. As a senior scientist, with 54 years of experience after the PhD., the topic was interesting simply because it asked the question ‘what really could be learned by and contributed by a novice, albeit a novice with professional experience.’

In fact, we learned that our respondents, who are also consumers, can easily grasp the features and uses of retail analytics, and that they reveal three distinct Mind-sets towards retail analytics. Mind-set 1: Focus on price/profitability; Mind-set 2: Focus on innovation/operations; and Mind-set 3: Focus on the outside world behavior. Each of these newly discovered mind-sets educates retail executives with knowledge never known to them before, guidance from consumers about how they might leverage retail analytics for their customers’ benefit and, perhaps, unleashing innovation benefitting both the shopper and the retailer. They can break from past tradition of analyzing from on high what they think is best for their shoppers.

The actual study was easy to implement. The use of the Idea Coach made things easier, although it was not clear to whether these were the ‘right questions’ to ask. The answers were not the issue. Rather, it was the questions, leading to the realization that emphasis is research should be put on asking an appropriate, meaningful question. So many of the AI generated questions seemed real until thought through, with the question of ‘what type of answer would this question engender’

The actual process itself revealed some unexpected benefits. The key benefit might be said to be ‘sharpening’ and improving the question, until one reaches a powerful question. The reality of this single powerful question is important. We are not taught important, seminal questions to ask. And all too often, when these questions are asked, their importance is often unrecognized.

To sum up, this paper demonstrates the increasing ease with which a novice can use computer technology to learn by exploration. The templated version of Mind Genomics makes it possible for the novice to move quickly from issue to empirical results. At the front-end AI provides a Socratic way of creating questions, that create an education on the topic. At the back end, the AI allows the researcher to understand the results from a variety of perspectives, provided by the different queries.

If we were to surmise the potential of the approach, we might find its greatest use in the world of education, to teach [8]. Teaching may not be limited to students, but rather ‘teaching’ here is used in the general sense, to instruct a person about a specific topic [9,10]. The unique combination of AI at the front and back, coupled with real human responses, provides a powerful tool to explore many dimensions of being, from the simplest to the profound, rapidly, iteratively, and with the potential of opening entirely new disciplines as the information from related studies is aggregated into a coherent whole with many different facets.

References

1. Oracle Corporation (2023) “What Is Retail Analytics? The Ultimate Guide.”
2. Kinsey MC & Co. (2021) “The value of getting personalization right—or wrong—is multiplying.”
3. Sprenger J (2011) Hypothetico‐deductive confirmation. Philosophy Compass, 6: 497-508.
4. Moskowitz HR (1971) The sweetness and pleasantness of sugars. The American Journal of Psychology, 84: 387-405.
5. Zemel R, Choudhuri SG, Gere A, Upreti H, Deitel Y, et al. (2019) Mind, Consumers, and Dairy: Applying Artificial Intelligence, Mind Genomics, and Predictive Viewpoint Typing. In Current Issues and Challenges in the Dairy Industry. Intech Open.
6. Gofman A, Moskowitz H (2010) Isomorphic permuted experimental designs and their application in conjoint analysis. Journal of Sensory Studies 25: 127-145.
7. Likas A, Vlassis N, Verbeek JJ (2003) The global k-means clustering algorithm. Pattern Recognition 36: 451-461.
8. Clancey WJ, Hoffman RR (2021) Methods and standards for research on explainable artificial intelligence: lessons from intelligent tutoring systems. Applied AI Letters 2:53.
9. Kim TW, Mejia S (2019) From artificial intelligence to artificial wisdom: what Socrates teaches us. Computer 52: 70-74.
10. Lara F, Deckers J (2020) Artificial intelligence as a socratic assistant for moral enhancement. Neuroethics 13: 275-287.

DOI: 10.31038/PSYJ.2023544

Abstract

54 respondents from an internet-based panel across the United States each evaluated uniquely different sets of 24 systematically varied ‘vignettes,’ (combinations of messages) about elderberry wine. The messages were created by artificial intelligence (Idea Coach), and afterwards combined into the vignettes according to an underlying experimental design which prescribed the appropriate combinations to use for subsequent regression analysis. Respondents rated each vignette using a two-dimensional rating scale, one dimension representing fit to the respondent (for me vs not for me), the second dimension representing understandability of the message. The data reveal three mind-sets. The study demonstrated the simplicity, speed, and economics of combining artificial intelligence, experimental design, and subsequent human evaluation. The output becomes a scalable bank of subjective information on a topic which is unfamiliar (elderberry wine), with this bank of information combining Socratic learning in a new topic coupled with feedback from real consumers about the information developed through artificial intelligence.

Introduction

The development of new products in the world of commerce is often costly, error-filled, and unduly long. Some of the issues may result from risk-avoidance, a phenomenon rampant in corporations, especially in slow-moving categories such as foods and beverages. When a company in electronics, for example, fails to avail itself of important technology to create new products that company is likely to suffer, often quickly, as its competitors rush to overtake it, doing so at hot speed. Not so in the world of food, even the world-of food start-ups, where the feeling is that there is not really much risk, that the competition moves slowly, and the technology is really not as valuable as the instincts and intuitions of the entrepreneur or the corporate president. The foregoing holds in classic, multi-layer multi-nationals as well as in the starts powered by the ingenue entrepreneur.

At the same time that the world of food development moves cautiously, there is an evolving world of speed, at almost any price. This world has emerged during the past decades due to the confluence of three factors, respectively the computer for processing, the internet for connection, and most recently artificial intelligence (AI) for rapid ‘thinking’ or at least rapid and seeming intelligent processing of text information in a way which seems intelligent. These three factors are making it possible to create ideas, test these ideas, and even expand them in what figuratively be an ‘eyeblink’ in the corporate timeline. What took hours, days, weeks, now can take minutes and seconds.

With the foregoing paragraphs as background, the author has begun a series of studies, small studies to be sure, on topics of daily life. The approach uses experimental design of ideas, mixtures of ideas presented to the respondent, the ratings of these mixtures revealing how each idea or message ‘drives’ the interest of the respondent. Using these tools of computer, internet, and now artificial intelligence, the author has pushed the study of ideas in the food industry down from a pedestal of scientific perfection to an act that even a grade school student can do, and even master after a moderate amount of practice [1-3].

The Mind Genomics Approach – Steps Towards Rapid Ideation

In order to demonstrate the power of new methods for product design, the author conducted a class experiment in April 2023, with students from the University of Florida in Gainesville. The approach was a DIY (do-it-yourself) approach for an advanced version of conjoint measurement, Mind Genomics. The specifics of Mind Genomics have been presented in detail in various papers published since the early days of the 21^st century. The reader is encouraged to look at the different topics covered. This paper will once again present the method, and the new development enabled by popular and available methods for artificial intelligence using the popular Chat GPT [4-6].

The ingoing, perhaps heretical and counterintuitive assumption, was that one could do a study within two hours, a study beginning with little or no knowledge about a field and emerge after those two hours with deep information about a topic. The topic chosen during the active initial back and forth was ‘elderberry wine,’ a wine of Asian origin (No et. al., 1980). The students who designed study had heard of elderberry wine, but were not familiar with the wine, making the exercise a challenge and enjoyable learning experience. The choice of elderberry wine emerges after about an unmoderated, 20-second class ‘discussion’ about ‘a topic, any topic having to do with foods.’

There is a modest-sized literature about elderberry wine, but a growing one, because of due to evolving consumer interest. At the same time, elderberry has received attention by horticulturists as well in part because of the increasing recognition of its health properties [7-11].

Table 1 presents the input information about the study. The table comes from a summarized report of the study automatically generated at the completion of the field work. The table provides the study title, date, purpose of the study as the researcher defines it, keywords for later sorting, self-profiling attitude questions, the respondent orientation (kept very simple), and the rating scale. All of this information is automatically incorporated into the Excel report.

The Mind Genomics process is templated, following choreographed sets which set up the experiment, run the actual experiment, and automatically analyze the data.

Table 1: Information about the study provided by the Excel report returned to the researcher at the end of the field work.

Introduce to the Process and Select the Topic (Elderberry Wine)

The exercise was set up so that the students would be introduced to the Mind Genomics process through a two-minute ‘elevator pitch’. The class was told that they would choose a topic, run a study, get results, and discuss the preliminary results. The students we unprepared, but as noted above, the decision was made to study elderberry wine. It is important to note that in no way was the topic to be focused on a so-called ‘burning issue’ or ‘hole in’ literature. The topic was selected almost randomly. It was at this point that the study had been registered as ‘elderberry wine’, the class as researchers filled out some checklists on using English as the language, and agreeing to not obtain information that could identify the respondents, except with the permission of the respondents.

Create Four Questions through Idea Coach

Mind Genomics works by a Socratic method, posing questions to obtain answers, combining the answers, and having respondents evaluate the combinations. Figure 1 shows the screen requesting the four questions (left panel), and the four questions actually selected (right panel).

Figure 1: The request for four questions dealing with elderberry wine (left panel), the Idea Coach for generating questions through AI (middle panel) and the four questions selected from the AI suggestions (right panel).

To generate the four questions in Table 1 is generally a function of one’s familiarity with the topic, and the predilection of the research group to come to an agreement. Often the group is unfamiliar with the topic, necessitating what ends up being interminable discussion and delay as the individuals in the group grapple with the appropriate questions to ask. The issue becomes even more vexing when the parties feel that they only have ‘one chance’ to run the experiment. It is at that point, the feeling of one-chance-only, that the participants in the research program end up ‘freezing’, often with the unhappy consequence that the project falls apart.

During the past several years users have continued to request, and eventually insist that the set-up of a Mind Genomics experiment be more ‘user-friendly.’ Almost all suggestions have included something about making it easier to generate questions, and to a lesser degree, to generate answers to the questions. As an aside, it appears that instead of trusting their own intuition and thinking, many individuals prefer ideas that have somehow been ‘vetted.’ This desire to have assistance in creating questions and answers, that assistance provided by an electronic ‘third party’, led to the creation of the Idea Coach. The Idea Coach is simply a set of AI prompts, based upon the ‘squib’ written about the topic. The squib is submitted to the AI program embedded in the setup, and generates 25-30 questions. Table 2 shows 60 unique questions emerging from four passes though the Idea Coach. With 30 questions there should have been 120 questions, but only 60 were different from each other.

Table 2: 60 unique questions emerging from four runs of the Idea Coach, each run returning a set of 30 questions about elderberry wine.

Create Four Answers to Each Question through Idea Coach

Idea Coach once again uses artificial intelligence to generate sets of 15 answers to each question. The researcher can select the answers of interest and edit them. Table 3 shows the answers for the first question (flavor profile) from three consecutive iterations of answers. In this case the 45 answers differ from each other. Figure 2 shows the screenshots of the final set of answers to the four questions selected.

Table 3: Three sets of answers to the first question (flavor profile)

 

Figure 2: Screen shots of the four answers to each question

The mechanism by which Idea Coach provides the questions and answers remains a trade secret of the company providing the AI system. What is important, however, is the rapid ‘learning’ by the Socratic method, question-and-answer, although here the learning might be in from parallel questions and answers, as the squib generates 25-30 questions, and the question generates 15 answers. One might happily imagine the potential of educating oneself on a topic such as elderberry, simply by two, three, four, or even five iterations of squib → 25-30 candidate questions → 15 candidate answers for each candidate question.

Create an Experimental Design Which Specifies the Combinations of Elements (Answers)

Rather than instructing the respondent to rate each of the 16 elements, one element at a time, the Mind Genomics approach presents combinations of these elements, short descriptions of the wine. Short descriptions are easier to judge because they tell a more complete story than do single elements. The respondent will evaluate a set of 24 vignettes, the aforementioned combinations.

In the experimental design element (viz., answer) appears five times in the 24 vignettes and is absent 19 times. No vignette comprises more than one element from any question. As a consequence, 20 of the 24 vignettes contain one answer from a question, whereas four vignettes of 24 are absent answers. Finally, across all 24 vignettes there are combinations comprising two answers, three answers, and finally four answers, but no vignette with only one answer. This specific design ensures that the researcher can analyze the data using standard statistical tools such as OLS (ordinary least squares) regression [12].

A unique, patented aspect of Mind Genomics is that each respondent evaluates a different set of combinations. The mathematical structure of the combination remains the same, but the specific combinations differ from one respondent to another. This difference is created by a permutation scheme described by [13]. The benefit of the permutation scheme is that the researcher need not know anything about the topic. The experiment allows the researcher to explore a great many combinations, analyze the data at the level of the individual respondent, and as a result uncover patterns that might not even have been imagined at the start of the study.

Create an Orientation and then Create the Rating Question that the Respondent Uses to Evaluate the Vignette. The Orientation is Simple: Read this Description about a New Elderberry Wine. How Do You Feel?

The rating scale actually comprises two dimensions. The first dimension is interest (Not for me vs For me). The second dimension is understandability (don’t get it versus get it). The two dimensions allow the researcher to understand the mind of the respondent more deeply, both in terms of emotion (for me / not for me) and intellect (get it / don’t get it).

Tell us what phrase best fits …after you read this.

1=Not for me … AND just don’t get it

2=Not for me … BUT I do get it

3=Can’t answer

4=For me … BUT just don’t get it

5=For me … AND… I do get it.

Create Self-profiling Questions

The Mind Genomics process enables the respondent to profile herself or himself on attitudes that would not be known from knowing who the response IS. At the start of the study the researcher can create up to eight such self-profiling questions. In addition, the Mind Genomics process automatically asks the respondent’s age and gender. The foregoing information (self-profiling as well as age and gender) are attached to the data provided by the respondent when rating the vignette. The self-profiling information will be used to create subgroups. Figure 3 shows an example of a self-profiling question. The Mind Genomics program allows the researcher to create up to eight such self-profiling questions, each with eight answers.

Figure 3: One of the self-profiling questions

‘Field the Study’ with Respondents

Figure 4 shows the information that the researcher provides to the Mind Genomics program. This information includes the number of respondents, how the respondents will be sourced, and whether or not the researcher wants to ‘privatize’ the respondent data. As of this writing (May 2023), the Mind Genomics platform is very low cost for the artificial intelligence (Idea Coach and summarization in the results). The researcher can either provide her or his own respondents at a low cost ($2/respondent for processing), use a third-party group ($2/respondent for processing), or use a Mind-Genomics approved supplier (Luc.id), with an approximate per respondent fee of $4-$6 for recruiting and processing. The researcher can make the data fully private for an extra $2/respondent. In this way the Mind Genomics fees can be kept very low, a boon to students who can explore a topic in depth, and actually run a small (or large study) on elements of interest, if desired.

Not shown is the set of screens which allow the researcher to specify country, age range, gender, education, income, children, and so forth for the respondent. The typical ‘field time’ to execute the experiment is about 60 minutes for 100-200 easy to find respondents.

Figure 4: The researcher specifies parameters about the field execution

Create the Database

Each record of the database corresponds to a vignette. The record has these columns:

Column 1 – Study name

Column 2 – Respondent identification number

Columns 3,4 – Age, Gender Columns 5-6 – Answers to the two self-profiling questions created by the researcher

Columns 8-23 – One column for each of th 16 elements. When the element is absent from the particular vignette, the cell has the value 0. When the element is present in the particular vignette the cell has the value 1. This is called ‘dummy coding’.

Column 24 – Test order of the vignette. Each respondent rated 24 vignettes, so the test order ranged from the first vignette tested (coded 01) to the last vignette tested (coded 24)

Column 25 – Rating assigned by the respondent to the vignette.

Column 26 – Response time in 100^ths of a second, defined as the elapsed time between the presentation of the vignette on the screen and the time that the respondent assigned the rating.

Column 27-30 – Create four new binary variables by a re-code of the rating to a binary valuer 0 or 100 (as well as the addition of a vanishingly small random number to the newly created binary variable)

Column 27 – Create variable ‘For Me.’ Ratings 5, 4 re-coded to 100, rating 3, 2,1 re-coded to 0.

Column 28 – Create variable ‘Not for Me’. Ratings 1, 2 re-coded to 100, 3,4,5 re-coded to 0.

Column 29 – Create variable ‘Get It’. Ratings 5,2 recorded to 100, rating 4, 3, 1 re-coded to 0.

Column 30 – Create variable ‘Don’t Get It’. Ratings 4,1 re-coded to 100. Ratings 5,3,2 re-coded to 0.

The database is set up for dummy variable regression analysis, either at the level of the individual respondent or at the level of the group. The original experimental design specified 24 different combinations, with the combinations being precisely those which ensure that each element appears equally often, and that each of the 16 elements are statistically independent.

The analysis will focus only on the ‘For Me’ ratings.

Create Equations (Models) Relating the Presence/Absence of Elements to Ratings

The equations are estimated using OLS (ordinary least squares) regression. For this analysis we express the equation in the standard way, using an additive constant:

Binary Variable = k₀ + k₁(A1) + k₂(A2) … k₁₆(D4)

The additive constant, k₀, shows the expected percent of responses to be assigned if there were no elements in the vignette. Of course, by design, all vignettes comprise a minimum of two and a maximum of four elements so that the additive constant should be considered a baseline.

Search for the Patterns

The patterns should emerge from the coefficients for the dependent variable ‘For Me’ (ratings 5 and 4 converted to 100, ratings 1,2 and 3 converted to 0). Table 4 shows the coefficients for total panel and for binary gender (male vs female). Mind Genomics returns with a great many coefficients. Table 4 and the remaining tables of coefficients (for mind-sets) show only the positive coefficients, viz, those 2 or higher.

Table 4: Performance of the elements by total and gender using R54 (For me) as the dependent variable. Only positive coefficients, 2 or higher, are shown.

The additive constant gives a sense of the percent of responses that would be 5 or 4 for vignettes that were empty, viz, absent elements. By design this is not possible, but the regression process can estimate this additive constant, which is typically considered to be a correction factor (Burton, 2021). Table 4 shows the additive constant to be about 46-47, suggesting that half of the time we can expect a rating of 4 or 5 for vignette or concept about elderberry wine, even when no elements to deeper detail.

Table 4 suggests that in terms of ‘interest’ (viz., For Me), a few elements perform strongly, and in fact elements that might not have been even thought of without the use of the AI-powered Idea Coach. These are ‘sensory testing’ for males, and ‘restaurant’ and ‘holographic labels’ for females. The benefit of creating elements with the assistance of AI is just this ‘out of the box’ thinking, with the researcher having the power to accept the suggestion or reject the suggestion by simple choice, and indeed to test the suggestion again in an easily run of the study with some new elements, new respondents.

Uncovering New-to-the-World Mind-Sets through Clustering

It is in the DNA of the scientific mind to look for basic causes, fundamentals of a situation. Although scientists and consumer researchers have attempted to develop profiles of archetypes, idealized profiles, these archetypes are too general, and fail to capture the granularity of everyday experience. Indeed, any attempt to divide people from the ‘top down’ is destined to fail because at the level of actual experience there are so many idiosyncratic factors that the archetypes simply do not have the ability to address [14].

The Mind Genomics approach works in the opposite direction, starting at the level of granular for a specific issue or situation, looking at the different dimensions of that granular situation, testing alternatives or expressions of each dimension, and then uncovering parallel groups of individuals or clusters for that situation. The clusters can be thought of as archetypes, not general ones, but archetypes of a specific situation, mind-sets in the language of Mind Genomics.

The statistics of Mind Genomics readily enable the researcher to discover these mind-sets, even without any ingoing knowledge. The approach simply creates individual level models of the type above shown for the total panel, or for any subgroup. Each individual generates a model, a model which is statistically valid because the 24 vignettes for each respondent had been created according to an underlying experimental design. The 24 vignettes are precisely arrayed to allow for OLS regression to be done on the data from each respondent. Each respondent produces 16 coefficients and an additive constant. Afterwards, the respondents are clustered by k-means clustering [15] first into two non-overlapping and exhaustive groups, (2-mindset solution), and then into three non-overlapping and exhaustive solutions (3-indset solution).

Clustering it follows purely mathematical criteria, e.g., minimize the sum of ‘distances’ between people in a cluster while at the same time maximize the distances between the centroids of the different clusters. It is left to the researcher to choose the number of clusters or mind-sets, and to name each cluster. Two good criteria are parsimony (fewer clusters are better), and interpretability (the clusters must tell a reasonably clear story).

For Mind Genomics studies, the measure of distance is the expression (1-Pearson Correlation). The Pearson Correlation coefficient measures the strength and nature of the linear relation between two sets of numbers, in our case the numbers coming from the 16 coefficients. The distance is small, viz., 0, when the Pearson correlation is +1 (1-1 = 0), occurring when the two sets of coefficients are parallel to each other. The distance is greatest, viz., 2 when the Pearson correlation is -1 (1 – 1 = 2), occurring when the two sets of coefficients go in precisely opposite directions.

Mind Genomics clustering usually reveals quite simple groups, the patterns often clear, ‘jumping out’ at the researcher. Table 5 shows the strong performing elements for the two and then the three mind-sets (abbreviated MS). The two mind-sets solution shows a very simple pattern, namely that which is familiar (venue for MS1 versus flavor for MS2). The three-mind-set solution is more intriguing, suggesting Label, Information, Venue, respectively. The three-mind-set solution is not perfect, since there are some strong-performing elements appealing to the mind-set slightly ‘off’ from the main interest of the mind-set.

Table 5 once again shows the ability of the OLS regression to uncover relevant coefficients, often coefficients which ‘make sense’ in their similarity to each other for a specific mind-set.

Table 5: Performance of the elements by total and both two and three mind-sets. gender using R54 (For me) as the dependent variable. Only positive coefficients, 2 or higher, are shown.

How Good are the Results?

Experienced researchers working in the world of inferential statistics and hypothesis testing measure their ‘performance’ by the likelihood that their hypothesis has not been falsified (Sprenger, 2011). The hypothetico deductive system of science is geared toward the creation, testing, acceptance/abandonment of hypothesis as the science moves slowly along. As famed scientist Max Planck opined ‘science advances one funeral at a time’ [16]. an experienced-based aphorism similar to the somewhat longer, more poetic but equally powerful idea from Tennyson’s Le morte d’Arthur “The old order changeth, yielding place to new, … Lest one good custom should corrupt the world “ (Sider, 2013).

With the evolution of research and its introduction into the world of education and application, the introductions often geared to ‘newbies’ (people without research experience), a common question is ‘how did we do?’. These newbies, students, others, do not have the wealth of experience, the years of data analysis, and the know-how about going to the extant ‘literature’ to compare their findings with what has been done. These newbies, aspiring researchers, need reinforcement about their work, e.g., a ‘score’ which tells them just how good their data are. In our over-measured society people use scores as an index of performance and a measure of growth.

One of the developments of Mind Genomics is the IDT, the index of divergent thought. The organizing principal underneath the IDT is that the positive coefficients, or more correctly the weighted positive coefficients, show how strongly the element ‘drives’ the rating. In other words, the IDT show how ‘on target’ the researcher has been by choosing elements to drive a dependent variable, that dependent variable here being ‘For Me.’

The IDT computations appear in Table 6. The total panel results account for 1/3 of the weight; the two mind-sets together account for 1/3 of the weight, and finally the three mind-sets together account for 1/3 of the weight. The stronger the performance of the coefficient for the total panel, the higher will be the IDT because that single high coefficient will in turn be multiplied by the value 0.33 for the total panel. In contrast, consider the value of that same high coefficient, but this time for MS 2 of 3, with 14 respondents, and a weight around 0.10. The contribution will be a lot lower. For this study the IDT is 46, reasonable. Unpublished values for the IDT in other studies have ranged from a high around to a low around 20.

Table 6: The IDT (Index of Divergent Thought)

Were the researcher to systematically vary aspects of the study and then measure the IDT for each aspect, the studies would move beyond informing about the world, and become a measure of the ‘impact’ of the different variables. Perhaps, most important for students is a measure of how well they understand the topic based upon the elements they select, the rating scale they use, the respondents they choose, and their experience as they iterate from one study to the next with clear human feedback. Much remains to be done.

Summarizing the Results Using Artificial Intelligence

The original objective of the study was to demonstrate the speed, power, and cost of Mind Genomics. As such, one of the goals was to see how quickly the key insights could be given to the reader in a format immediately ready for further efforts, including application or follow-on research. The first requirement was that the insight to be presented had to emerge in a robust way from the data, thus linking the insights to the actual experiment. The second requirement was that the insight had to be multifaceted, produced by clearly stated queries. The third requirement is that the insights had to be scalable, emerging from a few to many queries (many being > 10), with the insights emerging automatically. The effort stopped short of automatically creating a preliminary summarization document in the form of a ‘working paper’, but that next step is increasingly within reach. This first step to summarize the results used six queries provided to the AI program, with the instruction to look only at elements scoring +6 or higher for the subgroup.

The six summarization queries were submitted to the AI program, with the summarization done for each defined subgroup in the population, and done twice, once for the ‘TOP” (ratings 5,4 → 100), once for the ‘BOT’ (ratings 1,2→ 100). The defined subgroups were gender, age, response to the various questions in the self-defining questionnaire at the start of the study and finally to the two and the three mind-set solutions. Table 7 shows the summarization of the results for each of the three mind-sets for the TOP values (Rating 5,4, for ME).

Table 7: AI summarization of the three mind-sets

Discussion and Conclusions

The tradition of scientific research has become increasingly professionalized during the past centuries. What started out as the explorations of amateurs into a world hardly known has evolved into the world of science and academe that we know today., replete with societies, with journals, with the inevitable issue of who can publish what, and of course what exact constitutes publishable work. If that is not sufficient, the issues emerging involve the invisible networks of researchers who know each other and give each other help or in some unhappy cases just the opposite. And finally, there is the issue of funding research, funding publication and the need to survive the publish-or-perish world. In the words of an unnamed colleague, ‘we are all fighting for a sliver of the unpredictable funding pie.’

Within this world of discomfort and competitive behavior, the efforts of students, aspiring professionals, end up being crushed more often than not by an invisible college and rules of what makes science valid. All too often, the focus on being safe and correct ends up discouraging the researcher. Within this world, the Mind Genomics effort produces a system which expands the vision and hope of the amateur researcher, providing the potential of systematized, scientific, often even interesting exploration. It is within that spirit that this paper is presented, not so much as the convenient but hardly explored topic of elderberry wine as much as the exploration of what people just might do if given tools to empower their curiosity.

References

1. Mendoza C, Mendoza C, Deitel Y, Rappaport SD, Moskowitz HR (2023) A. Empowering young people to become researchers: What does it take to become a police officer? Psychology Journal Research Open 5(2): 1-12.
2. Mendoza C, Mendoza C, Deitel Y, Rappaport S, Moskowitz HR (2023) B, Empowering Young researchers through Mind Genomics: What will third grade mathematics look like in 10 years? Psychology Journal Research Open 5(3): 1-15.
3. Mendoza C, Mendoza C, Rappaport S, Deitel Y, Moskowitz H (2023) C, Empowering young researchers: Cognitive economics and the features associated with minimum wage. Ageing Science and Mental Health Studies 7(1): 1-9
4. Gere A, Radványi D, Moskowitz H (2019) Consumer Perspectives about innovations in traditional foods. In Innovations in Traditional Foods, 53: 84. Woodhead Publishing (2019).
5. Lieberman, L.E, Moskowitz, D.I, Moskowitz, H.R (2012) Consumer research in the wine industry: New applications of conjoint measurement. In: Alcoholic Beverages. 395- Woodhead Publishing.
6. Moskowitz, H.R. Gofman A. Beckley J, Ashman H (2006) Founding a new science: Mind genomics. Journal of Sensory Studies 21: 266-307.
7. Cernusca, M.M. Gold, M.A, 7 Godsey, L.D (2011) Elderberry Market Research.
8. Cernusca, M.M, Gold, M.A, Godsey, L.D (2012) Using the Porter model to analyze the US elderberry industry. Agroforest Systems 86: 365-377.
9. Byers, P.L, Thomas, A.L (2005) Elderberry research and production in Missouri. N Y Berry News 4: 11.
10. Charlebois D, Byers P, Finn, C.E, Thomas, A.L (2010) Elderberry: botany, horticulture, potential. Horticultural Review 37: 213-280
11. Finn CE, Thomas AL, Byers PL, Kemal M (2008) Evaluation of American Sambucus canadensis and European S. nigra elderberry genotypes grown in diverse environments and implications for cultivar development. Horticultural Science 435: 1385-1391.
12. Burton AL (2021) OLS Linear regression. The Encyclopedia of Research Methods in Criminology and Criminal Justice 2: 509-514.
13. Gofman A, Moskowitz H (2010) Isomorphic permuted experimental designs and their application in conjoint analysis. Journal of Sensory Studies 251: 127-145.
14. Lloyd S, Woodside AG (2013) Animals, archetypes, and advertising A3: The theory and the practice of customer brand symbolism. Journal of Marketing Management, 29: 5-25.
15. Likas A, Vlassis N, Verbeek JJ (2003) The global k-means clustering algorithm. Pattern Recognition 362: 451-461.
16. Azoulay, P.Graff-Zivin J, Uzzi, B. Wang, D. Williams H, Evans, J.A, Jin, G.Z, Lu, S.F, Jones, B.F, Börner K, Lakhani, K.R, 2018. Toward a more scientific science. Science, 361: 1194-1197. [crossref]

DOI: 10.31038/PSYJ.2023543

Abstract

Respondents ages 15-21 each rated 24 vignettes, combinations of messages about the experience of learning a second language in high school. The vignettes were created according to an experimental design, developed from four aspects (questions), each with four elements (answers to the questions). The vignettes comprised combinations of 2-4 elements, at most one element from each aspect, rating each vignette on a two-dimensional scale (Describes me vs Does not describe me; Leaves me with a good feeling when I read vs leaves me with a bad feeling when I read it). The elements were chosen by a high school sophomore, to represent how a young researcher would approach the topic, and how the researcher could work with respondents of approximately the same age. The approach reveals differences in personal experience with learning a second language, across genders, age groups, and across emotional response to the experience. A few elements emerged as important when the respondents were divided by WHO they were (self-profiling). Stronger, more insight-driving differences emerged when the respondents were clustered into Mind-Sets, based upon the patterns of their responses. The experiment suggests the potential for having young researchers study their contemporaries, using a templated approach, but with the contribution of artificial intelligence (Idea Coach) to help the process while still keeping the young researcher deeply involved, and in control.

Introduction

The study reported here on the feelings about learning a second language emerged from a discussion between authors Kornstein and Moskowitz about the system of education, especially regarding the student experience, specifically language education. There is significant literature on the different aspects involved in learning a second language in high school, most of practical nature. There are many facets of the issue, although most of the literature focuses on issues involving how to teach the language, especially today with on-line classes , what to teach and how to with student issues [1-7].

Although the literature does have relevant papers on language learning from the perspective of the student, a great number of these are geared to solving issues which emerge from difficulties experienced by the student . There is little in the way of simple but scientific understanding of the quotidian, everyday experience. Such understanding is left to literature, often autobiographical but just as often fiction written from the point of view of the student in the experiential ‘moment’ of learning the language [8-11].

It is to the disciplined study of one’s recollection of learning a second language that we turn to in this paper. Rather than focusing on the nature of problems, we look at the actual experience using the emerging science of Mind Genomics. The objective is to lay out the alternative aspects of what is experienced, see what the respondents choose as their experience, and their feelings towards that which is experienced. The Mind Genomics approach provides a new approach to augment existing approaches in education. Mind Genomics has already been used to explore education in third grade mathematics, not done by teachers or scientists looking into the experience, but rather from the mind of an eight-year-old researcher, looking out from her own experience, to forecast what might happen in a decade (Mendoza et. al., 2023B). The foregoing applications of Mind Genomics to the quotidian world, the world of the everyday, is just one of a number of papers appearing now, papers which demonstrate the ability of the student researcher to approach the world in rational, scientific manner [12].

The Mind Genomics Worldview of Daily Experience

Mind Genomics is an emerging science of everyday experience, areas of daily life that are overlooked because of the nature of their sheer ordinariness. We live, however, in the world of the everyday when we make our decisions. Indeed, most of our world runs reasonably smoothly because people recognize the regularities of nature, including the regularities of the world around as created by other people. At the same time, one could ask a scientist to list out the key factors in almost an experience, classify them, and then explain how people make decisions regarding these regularities. The level of lack of knowledge will amaze. We know at an intuitive level, or more correctly, we are probably able to surmise what are the key factors in everyday life. On the other hand, for a specific situation, e.g., when trying to sell something to another person, a few moments of being challenged about what exactly to say to make the sale quickly reveals glaring gaps in practical knowledge.

It is to refocus one; s attention on the science of the sheer ordinary that Mind Genomics was born in the 1980’s. A great deal of the literature at that time was emerging from laboratories of behavioral science, where the test subject was presented to specific test stimuli in an artificial situation, to understand one or another behavioral principle. It was in this situation that Skinner’s Behaviorism began [13], with principles of reinforcement and continued behavior demonstrated through unusual situations, e.g., pre-defined ‘schedules of reinforcement’ to illustrate specific aspects of behavior.

At the same time as Behaviorism was uncovering principles of behavior using artificially created situations, there was another set of developments, best referred to as consumer psychology or consumer behavior. A great deal of that topic was developed to study how people make real world decisions. The stimuli might be ordinary, or systematically varied, but the topic was the essentials of everyday life, the things, and the experiences relevant to people. The interest was in the rules governing the ordinary, the behavior of the person as an everyday consumer. The interest in consumer psychology was motivated by science and by business at the same time. The key, however, remains the sheer real nature of the focus. What do consumers really want? [14-16].

The Mind Genomics Approach – Explicating it through a Study on Education

An easy way to understand the approach, results and implications of Mind Genomics is through a study. Studies with Mind Genomics are easy to set up, quick and inexpensive to implement, and rich with data revealing patterns, some already known, others delightfully new. The study presented here deals with the experience of students studying a second language in high school. The approach is to let a high school student be the researcher, let high school age and slightly older students be respondents, and uncover the mind of students perhaps not before well understood.

When developing the idea for this study on education, the selection of the ideas was specifically left to author Kornstein, a second-year high school student. Rather than imposing one’s ideas on the research, the decision was made to avoid any input except for technical (grammatical) changes of the raw materials (questions and answers/elements). The only outside input to the study was the creation of the two-sided rating scale, explained below. The strategy of making the study reflect the mind and interests of the high school student provides a new way to understand a topic, an understanding not so much from the outside in (adult studying the student), but rather from the inside studying the inside (student studying the student).

Step 1: Select the Name of the Study

Although one might consider naming to be minor, that is not the case. From many Mind Genomics studies the observation has continued to emerge that ‘naming’ the study is an important first step in focusing the researcher on the topic. All too often the first attempt to name the study ends up with the student providing an entire sentence about what is to be studied. Naming the study forces the student to become more open, not to focus on specifics. Parenthetically, the same issue occurs in today’s PhD researchers, who define themselves, their ‘field of study’ by the method that they use to gather the data, rather than by the science to which they are trying to contribute. These students and often newly minted researchers think of their science as their research method rather than the underlying research problem.

Figure 1 (top row, left panel) shows a screen shot of the BimiLeap program (www.BimiLeap.com), the program which allows the researcher to do the study by following a template.

Figure 1: Set up for the Mind Genomics study on studying a second language

Step 2 is the hardest part of the study for the researcher, often for the simple reason that we are not taught to ask questions, but rather taught to answer questions that are posed to us. Therefore, our thinking ends up being scattered. We can ‘tell a story’ if we are asked to do so, but we usually don’t think of a topic in terms of a story which unfolds, within a structure. Were we to be educated to do so, we might begin investigations with a series of questions allowing us to paint a simplified picture of a topic. Mind genomics works in that way.

In previous versions of Mind Genomics, the effort to create these ‘questions’ was so great that quite often the aspiring user simply ‘gave up,’ with a statement of resignation about simply not being able to create thee questions. Other researchers kept going, and in most cases afterwards successfully developed questions, after what seems in retrospect to have been a harrowing, frustrating experience.

Since the end of 2022 the Mind Genomics program, BimiLeap, has incorporated the Idea Coach, using artificial intelligence. The researcher simply types in something about the topic, preferably that ‘something’ being close to the specifics. Figure 1 (top row, right panel) shows the ‘squib’ or little paragraph describing the issue. Figure 1 (bottom row, left panel) shows some of the 30 questions emerging from Idea Coach. Figure 1 (bottom row, right panel) shows the four questions selected by the researcher.

A sense of a set of 30 questions emerging from Idea Coach appears in Table 1. Using Idea Coach with the same ‘squib; (top row, right panel) a second and third time will produce a different set of questions. Table 2, in turn, shows two sets of 15 answers to question 1, these answers produced by the same query to the Idea Coach, with Idea Coach returning with two different sets of answers.

Table 1: 30 questions emerging from Idea Coach when presented with the statement: What are good questions to ask about learning a second language in high school.

Table 2: Two sets of 15 answers each for question 1 (what is hard for me when learning a second language).

The actual questions and answers appear in Table 3. These are provided by the researcher, or provided by the Idea Coach, with the researcher enabled to modify them.

Table 3: The four questions and the four answers (elements) to each question as selected by the researcher and used in the study.

The actual test stimuli, however, moves beyond the single elements, and into combinations of elements. The combinations themselves are not random, although to an untrained eye they might appear to be a ‘blooming, buzzing confusion’ in the words of Harvard psychologist, William James [17]. Nothing could be further from the truth. The combinations or vignettes are put together using an underlying experimental design, a set of planned combinations, with the combinations set up to allow further statistical analyses. Figure 2 shows a screen shot of some combinations tested, that screen shot coming from the drop-down menu of the actual study, available to the researcher after the study is completed.

Each respondent evaluates a set of 24 vignettes, the vignettes set up with the following properties:

1. The 16 elements are statistically independent of each other. This statistical independence means that the data from each respondent can be analyzed by OLS (ordinary least-squares) regression modeling [18].
2. Each element appears exactly five times in the 24 vignettes evaluated by a respondent and is absent from 19 of the vignettes.
3. Each vignette comprises at most one element or answer from a question, but many vignettes comprise two or three elements, not four. This property, incompleteness, allows the researcher to use OLS regression to estimate the absolute contribution of each element to the rating or to a transformed rating, as will be discussed below.
4. Each respondent evaluates a different set of combinations, but mathematically the combinations evaluated by the respondent are ‘formally’ identical. That is, the system creates one basic design, the kernel design, and permutes the design so that each respondent ends up evaluating different combinations, but the design is of the same structure. That mathematics is powerful, because now everyone can be separately analyzed. Furthermore, and metaphorically like MRI (magnetic resonance imaging), the researcher can explore different combinations, and will end up with a better ‘picture’ of responses to the underlying elements. In other words, the researcher can explore the ideas, rather than simply waste money ‘testing’ the correctness of an idea. This latter thinking, exploration rather than confirmation, is a potential ‘game changing’ notion for psychological science [19].

Figure 2: A screen shot of four vignettes

Step 3: Create Self-profiling Questions, Create the Rating Scale, Record One’s Own Thoughts a about the Study, and Select the Source of Respondents for the Study, When the Study is Executed ‘in the Field’

Table 4 (Part A) presents these questions. They provide information about the respondent that would not be otherwise obtainable, since the identity of the respondent is confidential, maintained so by the on-line panel provider (Luc.id Inc.).

Table 4: Key information about the reported back as part of the results of the study. The table summarizes key features of the study, providing a record for the researcher.

Table 4 (Part B) presents the actual rating scale. The rating scale comprises two parts, a section dealing with whether the respondent feels that the test vignette describes the respondent (rating 5 and 4) or does not describe the respondent (rating 1 and 2). The second part of the scale deals with the feeling of the respondents about what is read, whether the feeling of a positive experience (ratings 5 and 2) or feeling of a negative experience (ratings 4 and 1).

Figure 3 shows four screen shots for different parts of the set up.

1. The top left panel shows a screen shot for one of the self-=profiling questions. The researcher needs simply to type the question in the large rectangle above, and type alternative answers in the small rectangles below.
2. The top right panel shows the open-ended question.
3. The bottom left panel shows the space where the researcher must provide some information about the study from the researcher’s own point of view. The information is often important to record at the start of the study, just after the study has been set up, so that the background for the research is ‘fresh’ in the researcher’s mind.
4. The bottom right panel shows the options about getting respondents, and about privatizing the data so no one can see the results.

Figure 3: Questions, final thoughts written by the researcher as a record, and choice of respondents

Step 4: Executing the Study with Respondents

The respondents are invited to participate by a company specializing in so-called on-line panels, viz., individuals who have agreed to participate in these studies The respondents are compensated by the panel provider. The identity of the respondent is unknown. The Mind Genomics researcher can specify aspects of the respondent, such gender, age, location (country, state), income and so forth. These selection criteria are built into the Mind Genomics system. Other features of recruiting can be done but require additional effort.

For this study, the objective was to work with students of high school age, and recent graduates, with a low of 14 years old and a high of 19 years old. The actual study returned 104 respondents, but 16 were eliminated from the database because they were either too old or too young, leaving 88 qualified respondents. The age was known because the respondent had to select the year of birth.

The study set up took about 90 minutes. The actual study in the ‘field’ was about 60 minutes from invitation to the respondent by the panel provider (Luc.id) to the completion of the study with 104 respondents. It is worth noting that many organizations prefer to use their own panelists which make sense, but which end up taking days, weeks, and sometimes never completes. With on-line panel providers involved, the weeks and days shrink to hours and minutes.

Create ‘Meaningful’ New Dependent Variables by Transforming the Rating

It is the nature of researchers to want to work with the numerical response scales that they use. After all, goes the thinking, the scale has been set up with a great deal of effort. Furthermore, in the mind of the researcher the scale is even more attractive because the ratings can be readily analyzed by the powerful statistical programs to which researchers have become accustomed.

There is only one negative to the foregoing pictures, something that practitioners learn, often painfully. That negative is the demonstration in action that the ‘client’ who will use the information, usually a manager, cannot easily interpret the data. No matter how one proudly proclaims the power of the scale, the bottom line is that many managers ask a simple, almost naïve question, the general sense being ‘what does this number mean? Is it good or bad? Should I worry?’

The foregoing question ‘what does this number mean’ is not to be sneered at. The question is serious because often the manager has to make a business decision based upon the data. The decision IS NOT embedded in an appeal to statistical significance. Users of data don’t understand that type of talk, not really. What they do understand is ‘Am I ok or am I in trouble.’ Maybe the words differ from person to person, but the essence of the question of ‘what do I do with these numbers?’

Over decade of experience implementing these studies and sharing/explaining the results, researchers have often ended up saying ‘good or bad’. It is much easier for a user of the data to understand good and bad as scale anchors and perhaps a percentage of the population choosing one or other, good vs bad, pass vs fail, promising vs not promising.

When setting up the scale, the researchers chose to combine two dimensions into one scale, and to explore the opposites of each dimension. The first dimension is ‘fits me’ versus doesn’t fit me’. Rather than focusing on graduation of ‘fit’ vs ‘doesn’t fit, we look at a yes/no response, namely does fit (5 or 4) or does not fit (1 or 2). The second dimension is how I feel when I read the description. Ratings 1 4 are ‘feel bad’, ratings 2 and 4 are ‘feel good.’ Scale point 3 is a catch-all for the inability of the respondent to decide.

One might think that this use of two scales is difficult for the respondent. That supposition is correct, but after the first few evaluations the respondent feels comfortable. The benefit of using the two scales emerges when the researcher can use different types of dependent variables, one having to do with the degree to which the respondent ‘identifies’ with the messages, the others having to do with the emotions generated by the messages, viz., positive versus negative.

Relate the Elements to the Responses Using Regression Modeling

The essence of Mind Genomics is to assign numbers to ideas, viz., to messages, with these numbers reflecting how the respondent thinks about the specific message. Rather than asking the respondents to rate the separate ideas using a scale, Mind Genomics approaches the task in a more natural way, one less prone to bias, far less prone to guesswork by the respondent. The respondent evaluates combinations of messages, rating each combination on a scale. The respondent simply knows the criterion for the rating (e.g., applies to me vs does not apply to me, or gives me a warm feeling versus a bad feeling).

The Mind Genomics system cannot be ‘gamed’. The respondent cannot guess the correct answer. Within one or two vignettes the respondent stops trying to intellectualize the process, settling down to a simple S-R behavior, stimulus-response. The respondent receives a vignette, and almost without thinking, the respondent ends up answering using the scale. One might think that the results would be a meaningless jumble, making no sense, but the results make a great deal of sense as we see below. The objective to measure the real feeling of the respondent is in sight.

The actual analysis is done by means of statistical analysis, specifically ‘OLS’ (ordinary least squares) regression (Alma, 2011), a widely available, easily to use, easy to understand statistical procedure. The independent variables are the 16 elements, the data are set or sets of 24 rows of data. The 24 rows ‘coded’ the presence/absence of the elements. A ‘1’ in a row for a specific element means that the element is present in that vignette. In contrast, a ‘0’ in a row for that specific element means that the element is absent.

The regression procedure estimates the 16 coefficients. We do not use an additive constant. Thus, the story is entirely within the pattern of the 16 coefficients. The output is an equation of the form: DV (dependent variable) = k₁(A1) + k₂(A2) … + k₁₆(D4)

Patterns of Responses

Tables 5-7 present the summarized data for the self-described groups. These groups are gender, age, effort, and the joy in learning (lowest versus highest). The strong performing elements (coefficients of 20 or higher) are shown in the shaded cells. Mind Genomics studies produce a great deal of data (viz., 16 coefficients for each group), requiring a strategy to present data in a way easy for the user to ‘get the picture.’ For this project, a coefficient +20 is significant in a statistical sense and seems to be meaningful from inspections of coefficients from previous experiments.

Table 5: Coefficients for models showing the degree to which each element is judged to describe the respondent assigning the rating. Each column refers to a specific subgroup of respondents, as the respondents describe themselves.

Table 6: Coefficients for models showing the degree to which each element is judged to generate a GOOD FEELING by the respondent. Each column refers to a specific subgroup of respondents, as the respondents describe themselves.

Table 7: Coefficients for models showing the degree to which each element is judged to generate a BAD FEELING by the respondent. Each column refers to a specific subgroup of respondents, as the respondents describe themselves.

Describes Me

Table 5 shows the elements which the respondent felt to describe them (ratings 5 and 4), whether the elements described a positive experience (rating 5) or a negative experience (rating 4). Strong performing elements appear in almost all self-defined groups. Overall, the three elements with which the respondents most frequently identified are:

A4                  Hard for me: Memorizing verb conjugations

C3                  Helps me: Songs in the target language

C1                  Helps me: Online courses

It is important to keep in mind that the data are a snapshot of a person’s mind. There are no hypotheses in these studies, although they are conjectures that might be substantiated. Mind Genomics is set up as a system to explore the way people think about experience, rather than as a system to prove or disprove a hypothesis.

I have a good feeling I read this vignette,

A Good Feeling after Reading the Vignette (R52)

Table 6 shows the same type of analysis, this time for those scale points which reflect the respondent’s rating of having a good feeling after reading the vignette, whether or not the respondent identifies with the vignette. Again, the strong performing elements are shown in shaded cells. All cells with coefficients of 0 or lower are left blank in order to allow the pattern to emerge more clearly.

The patterns which emerge are less clear. The first pattern is the absence of clearly strong performing elements for total panel, for genders, and for ages. The second pattern is the emergence of strong performing elements among those respondents who say that they found it hard to learn a second language. The third pattern, and the one potentially most instructive, is the one which emerges when we look at the responses of individuals who say that they were very unhappy when they learned a second language. There were three ways to learn language that these respondents felt gave them a good feeling when they read them:

D1          Learn best through: Structured lessons                                                                                  21

B4          I like to learn by: Listening                                                                                                        21

D3          Learn best through: Combination of structured lessons and creative activities                  21

A Bad Feeling after Reading the Vignette (R14)

Table 7 shows the same type of analysis, this time for those elements which reflect the respondent’s rating of having a bad feeling after reading the vignette, whether or not the respondent identifies with the vignette. Once again, we are confronted with some paradoxical results. The strong negative feelings emerge for four elements, and, paradoxically, only among those respondents who felt that they were very happy when learning the second language. These strong, negative feelings emerged for

C4        Helps me: Grammar workbooks                                      30

C1        Helps me: Online courses                                                 20

C3        Helps me: Songs in the target language                       26

A1         Hard for me: Memorizing vocabulary                           20

Deeper Understanding by Uncovering Mind-sets

A hallmark of Mind Genomics is the focus on mind-sets, defined operationally as groups of individuals who differ from each other in clear ways, when they are deciding about everyday issues or activities. In other words, different ways of thinking about the world of everyday. Mind Genomics moves beyond dividing people by WHO they are, or WHAT they do, or even how they THINK about general topics, focusing instead of the granular aspects of everyday life.

The project reported here on the way people think about learning a second language is set up for the discovery of mind-sets using the Mind Genomics methods. The actual process to discover the mind-sets has already been templated and is incorporated into the BimiLeap program. The only difference is that the ‘go-forward’ approach for Mind Genomics is to estimate regression models without the additive constant.

1. Using OLS (ordinary least-squares) regression, estimate the 16 coefficients for an equation on a respondent-by-respondent basis. Even though each of the 88 respondents in this study ended up with a different set of 24 vignettes, the underlying experimental design ensured that those 24 vignettes would be exactly the 24 needed for a valid regression model, with all 16 predictor variables (presence/absence of elements) statistically independent of each, with equal numbers of appearance (n=5) for each element, and the presence of ‘zero’ conditions, to allow the estimation of absolute values for the 16 coefficients
2. With the database of 88 rows of coefficients, use k-means clustering to generate exactly two groups, those groups defined by the ‘distance’ metric (1-Pearson R). The clustering was instructed to create two groups, whether the groups were interpretable or not. By definition, the clusters satisfied the appropriate mathematical criteria, viz., that according to the distance criterion, the respondents in a cluster were close together (minimize the distance between pairs of respondents), whereas the average profile of the 16 coefficients for the two clusters were as different as possible. This is the k-means clustering routine [20-23]. The results are two clusters, known as ‘mind-sets,’ in the language of Mind Genomics.
3. Table 8 shows the pairs of mind-sets for the three important dependent variables: R54 (Describes Me), R52 (Good Feeling after reading the vignette), and R41 (Bad Feeling after reading the vignette. Once again, the strong performing elements are shaded (coefficient of 20 or higher), and very low coefficients (1 or lower) and shown by empty cells.

Table 8: Coefficients for pairs of complementary mind-sets (clusters)

The creation of mind-sets allows radically different groups to emerge.

Describes Me (Rating 54) reveals that Mind=Set A describes themselves as learning through active participation, whereas Mind-Set B describes themselves as learning by listening and doing grammar exercises. A three-cluster solution (not shown) shows the same overlap of features describing how the mind=sets see themselves. It may well be that students do not really know how they best learn.

Gives me a good feeling (Rating 52) reveals that Mind-Set C feels best with traditional methods, whereas Mind-Set D feels best with creative activities woven into the learning process.

Gives me a BAD feeling (Rating 14) reveals that Mind-Set E feels worst with traditional methods, whereas Mind-Set F feels worst with on-line courses.

Discussion and Conclusions

Our knowledge about the experience of learning a second language in high school typically comes from adults, either teachers who observe the process of a student being educated, or from professionals who interview the student, filter/digest the information, and finally interpret and report what they heard. The literature is large, as one might expect, because of the singular importance of education to our society.

This study provides a new direction for understanding education, as well as other topics. The study is titled ‘Empowering young researchers: Exploration of teen responses to learning a second language’. The objective is to let the student be the researcher, select the topics to be researched, and use the Mind Genomics tool to streamline the research process, converting into learning, rather than onerous data preparation.. The template makes the research easy to do, allowing the student research to focus on the topic, and not be intimated by difficulties either in starting the process, doing the actual research, and analyzing the results.

With the above taken into consideration, it becomes clear that learning a second language is not a simple thing, not a ‘cut and dried’ process. Although messages are clear, and although mind-sets emerge, it is clear that even to a respondent presented with vignettes, there is no clear division of respondents into different mind-sets. The experience of learning a language appears to be amorphous, fluid, not something which is clear. The clarity of thinking and polarization of mind-sets revealed by student researchers using Mind Genomics does not emerge when we deal with high school and slightly older respondents rating vignettes about learning a second language. Whether that is a failure of the method, a possibility, or indication of a far more complex word needing significantly expanded efforts, remains to be seen. Clearly, however, people do identify with different ways of learning the language and do have memories. It is simply the overall patterns which may elude us. Mind Genomics makes it possible for everyone to learn and discover, whether professional, student, even interested layperson.

References

1. Fathman AK, Kessler (1992) Cooperative language learning in school contexts. Annual Review of Applied Linguistics, 13: 127-140.
2. Jones J, Coffey (2016) Modern Foreign Languages 5: 11: A Guide For Teachers. Taylor & Francis.
3. Pachler N, Redondo A (2007) A Practical Guide to Teaching Modern Foreign Languages In the Secondary School. Routledge.
4. Tochon FV (2015) Mobile experiences of an adolescent learning Spanish online in a twenty-first century high school. International Journal of Pedagogies and Learning 10: 91-106.
5. Bailey P, Onwuegbuzie AJ, Daley CE (2000) Correlates of anxiety at three stages of the foreign language learning process. Journal of language and Social Psychology 19: 474-490.
6. McCafferty SG (2006) Cooperative Learning and Second Language Teaching. Cambridge University Press.
7. Noels KA (2003) Learning Spanish as a second language: Learners’ orientations and perceptions of their teachers’ communication style. Language Learning 53: 97-136
8. Ehrman ME (1996) Understanding Second Language Learning Difficulties. Sage.
9. Raja B, Selvi K (2011) Causes of problems in learning English as a second language as perceived by higher secondary school students. Journal of English Language Teaching 1: 40-45.
10. Trang TTT, Baldauf Jr. RB, Moni K (2013) Investigating the development of foreign language anxiety: An autobiographical approach. Journal of Multilingual and Multicultural Development 34: 709-726.
11. Porras González NI (2010) Teaching English through stories: A meaningful and fun way for children to learn the language. Profile Issues in Teacher’s Professional Development 12: 95-106.
12. Mendoza C, Mendoza C, Rappaport S, Deitel Y, Moskowitz H (2023) (C) Empowering young researchers: Cognitive economics and the features associated with minimum wage. Ageing Science and Mental Health Studies 7(1): 1-9.
13. Smith LD, Woodward WR (1996) B.F. Skinner and behaviorism in American culture. Associated University Presses.
14. Moskowitz HR, Gofman A, Beckley J, Ashman H (2006) Founding a new science: Mind Genomics. Journal of Sensory Studies 21: 266-307.
15. Moskowitz HR (2012) Mind Genomics: The experimental inductive science of the ordinary and its application to aspects of food and feeding. Physiology & Behaviour 107: 606-613. [crossref]
16. Porretta S, Gere A, Radványi D, Moskowitz H (2019) Mind Genomics (Conjoint Analysis) The new concept research in the analysis of consumer behaviour and choice. Trends in Food Science & Technology 84: 29-33.
17. Dember WN (1990) William James on Sensation and Perception. Psychological Science 1: 163-166.
18. Alma ÖG (2011) Comparison of robust regression methods in linear regression. International Journal of Contemporary Mathematical Sciences 6: 409-421.
19. Gofman A, Moskowitz H (2010) Isomorphic permuted experimental designs and their application in conjoint analysis. Journal of Sensory Studies 25: 127-145.
20. Likas A, Vlassis N, Verbeek JJ (2003) The global k-means clustering algorithm. Pattern Recognition 36: 451-461.
21. Kissau SP, Kolano LQ, Wang C (2010) Perceptions of gender differences in high school students’ motivation to learn Spanish. Foreign Language Annals 43: 703-721.
22. Mendoza C, Mendoza C, Deitel Y, Rappaport SD, Moskowitz H (2023) (A) Empowering young people to become researchers: What does it take to become a police officer? Psychology Journal Research Open 5(2): 1-12.
23. Mendoza C, Mendoza C, Deitel Y, Rappaport S, Moskowitz HR (2023) Empowering young researchers through Mind Genomics: What will third grade mathematics look like in 10 years? Psychology Journal Research Open 5(3): 1-15.