CHAPTER 1
What are the basics you need to know to interpret statistical analysis results?
- Hypothesis, significance level, rejection region, test statistic, p-value -

Chapter 2
Before launching a new product, we conducted a marketing campaign where we asked half of the customers who visited the store if they would like to purchase the product, and provided the remaining customers with samples of the new product and then asked them if they would like to purchase it.
The average purchase intentions of customers who received samples differ from those who did not. Can we verify whether this difference is statistically significant?
- Independent samples t-test, dependent samples t-test, analysis of variance (ANOVA) -

CHAPTER 3
To demonstrate the effectiveness of the training, job expertise was measured for both the trained and non-trained groups.
The independent-samples t-test results showed no significant difference in the average expertise between the two groups. Does this mean the training was ineffective?
- Experimental design, analysis of covariance (ANCOVA) -

Chapter 4
I would like to understand the relationship between employee performance and various factors that are believed to influence it.
Is it possible to statistically determine whether each factor has a positive or negative impact on performance? 63
- Regression analysis -

CHAPTER 5
I would like to know what factors influence job satisfaction.
Can I perform a regression analysis by setting all available variables as independent variables? 85
- Theoretical background -

Chapter 6
Five independent variables were input into the regression model to predict performance based on theoretical background.
The analysis results showed that the independent variable, commuting time, had a positive effect on performance.
It is difficult to accept the interpretation that longer commuting times lead to higher performance.
What could have gone wrong? 95
- Multicollinearity -

CHAPTER 7
How do I know which of the performance-influencing factors has the greatest impact? 103
- Standardized regression coefficient -

CHAPTER 8
We divided the team leader's leadership style into four types.
Can you figure out how a team leader's leadership style affects subordinates' performance? 119
- Coding of categorical variables -

CHAPTER 9
Can you confirm whether the effect of welfare system satisfaction on job satisfaction varies by job level? 143
- Comparison of coefficients between groups -

Chapter 10
Can you confirm whether the effect of welfare system satisfaction on job satisfaction varies depending on the level of organizational adaptability? 157
- Moderation effect -

CHAPTER 11
We are operating a certification acquisition training course for key talents.
Is it possible to predict with probability whether each key talent will obtain a certification? 173
- Binomial logistic regression -

CHAPTER 12
Based on the customer's budget and information about the products they actually purchased, I want to predict which product they will purchase: A, B, or C.
197
- Multinomial logistic regression -

CHAPTER 13
Recently, many of our key talents have been losing their positions to competitors.
Is it possible to predict whether individual key talent will move on, and if so, when they're most likely to do so? 213
- Discrete-time survival analysis -

CHAPTER 14
jamovi 251

Conclusion 271
Bibliography 273
Index 276

Entering

Major League Baseball teams invest based on player performance data and analysis thereof, and Google establishes the basis for all human decision-making, centered around a department dedicated to People Analytics.
The media is constantly talking about the 4th Industrial Revolution's underlying technologies, such as artificial intelligence (AI), big data, and the Internet of Things (IoT), and the new changes they will bring.
On one hand, changes are beginning to appear in society, signaling the emergence of a generation that pursues rational individualism, following a generation steeped in traditional collectivist values.
Concerns are deepening about how to cooperate and adapt to the new generation, which has unique characteristics compared to previous generations.
This new wave of change also raises questions about the way we make decisions about people.
It demands a basis for traditional 'intuitive' decision-making.
The recent changes in business, technology, and society listed above encapsulate why people analytics is necessary and, furthermore, why this book was written.

Why People Analytics?

Moneyball (2011) is a film based on a true story from 2002.
Billy Beane (Brad Pitt) in Moneyball uses individual player records to identify players with a high ability to advance to first base.
Considering each individual's salary, we recruit players with the greatest cost-effectiveness and let go of players with lower salaries.
Many people ridiculed this unprecedented way of running a club.
But the result? A 20-game winning streak for Oakland, the weakest team in the league.
The team was composed of individuals who had a high probability of achieving the team's desired outcome of "advancing to first base," which led to the team's achievement of "team victory."

Recently, in the management field, attempts are being made to identify and control factors that influence results based on data about people.
Representative examples include companies such as Google and Microsoft.
Global consulting firms such as McKinsey are also conducting consulting based on data analysis.

For example, in the field of HR, the probability of individual talents changing jobs is also predicted.
Beyond simple predictions, we aim to uncover the factors that must be managed to prevent the loss of essential talent within a company.
If we can identify the factors that influence the turnover rate, we can devise ways to control these factors and thereby induce retention of key talent.
The presence of key talent within a company that enables core technologies is a critical factor in determining a company's success or failure, just like hardware such as production facilities or research facilities.

Let's take another example.
In the field of marketing, factors such as customer preference for a company's products, time spent visiting a store, and friendliness of store employees are measured to predict whether or not a customer will actually purchase a product.
Well-made products and services are important, but without research on the customers who consume and use them, they will simply become unsellable products and services.
Areas like R&D and production that make a company's products and services possible are still important, but now research and decision-making about 'people'—the 'employees' who make R&D and production possible, and the 'customers' who will purchase our products and services—have also become important.

How have we been making decisions about these people all along? It's difficult to say for sure, but many corporate HR and marketing departments, which are tasked with making decisions about people, may have been relying on "experience" and "intuition" developed from that experience, rather than analyzing data.
Some might ask:
The experience, intuition, and instinct that we have accumulated so far have created the success stories of small and medium-sized businesses and large corporations. Will it be a problem to continue doing the same in the future?
As mentioned earlier, a new era is coming.
We live in an era where new technologies are coming in and we are working with a new generation.
Just as the back wave of the Yangtze River pushes the front wave, change has become the norm, and its speed is unprecedented.
The successful experiences of 20 to 30 years ago and the intuition and instincts developed from those experiences do not guarantee the same success as in the past in this new era where it is impossible to predict even an inch ahead.
Our decision-making methods must also change accordingly, especially when it comes to people in this new era.
This is why you need People Analytics.

Why do you need this book?

The importance of people analytics has recently been emphasized, and leading companies at home and abroad are making efforts to introduce it in an advanced manner.
In line with this, books are being published that contain various examples of companies applying People Analytics to solve problems or attempt new initiatives.
The problem is that while these "introductions to new approaches" continue, it's difficult to find books that actually address how to measure recruitment, collect data, and analyze that data.
Conversely, statistics-focused books that address the "how" are not easy for business managers with limited quantitative background to understand.
Another problem is that it is not friendly to how to actually run statistical programs and interpret the results of the analysis.
In this context, I wrote this book with a sense of responsibility as a scholar specializing in behavioral science measurement and statistics, who wants to help those in need by providing better methods than before.
Each chapter of ?People Analytics? is designed to help business managers who want to implement People Analytics through the following structure and approach.

Chapter / Situation

Each chapter is organized based on the questions and topics that corporate managers actually wonder about when applying People Analytics.
The questions themselves were organized as the titles of each chapter, and the first component of each chapter, 'Situation', specifically described and presented the questioner's situation.
If you do not have basic knowledge of statistics, it is recommended to read the book in order. However, if you have some background knowledge and the problem you need to solve matches the topic of each chapter, you can also read excerpts from that chapter.

Solution / Statistics

The second component, 'Solution', presents and briefly explains the statistical methodology that must be applied to solve the topic presented in each chapter's Situation.
The third component, 'Statistics', supplements the theoretical explanation of the statistical methodology presented in the solution.
Because statistical methodology is a comprehensive result of mathematics and logic, the barrier to entry is high.
If you've ever tried to learn statistics before and given up halfway through, it's probably because you had trouble understanding the complex formulas and various terms.
In this book, formulas are used sparingly and only where necessary for these readers.
If you have a basic level of mathematical knowledge, I don't think you will have much difficulty understanding the formulas.
Likewise, with regard to statistical terminology, only absolutely necessary terms were used.
Additionally, we tried to increase understanding by continuously giving examples and repeatedly explaining.

Application/Same application in different situations

In 'Application', we covered how to manipulate actual statistical programs to solve the situations presented in the Situation.
It is structured so that each step can be followed step by step through screen capture.
Additionally, we have increased the applicability of the analysis results by providing detailed instructions on how to interpret them.
For chapters with practice, we recommend that you download the practice files for that chapter and analyze them as in the book, and then apply them to your own materials.
As with any other academic discipline, data analysis methods cannot be fully learned unless they are applied.
In 'Same application in different situations', additional topics that can be solved by applying the same measurement and statistical methods are presented.
I hope that you will continue to think about situations where you can apply what you have learned through this, and increase your understanding and applicability.
* Data download url: Provided on the DATA page of http://www.seheehong.com

In the first chapter of this book, the basic knowledge and concepts necessary to understand measurement and statistics are presented.
Additionally, the last chapter contains an introduction to the programs used in the practice, as well as instructions for downloading, installing, and configuring them.
In order to understand data analysis methods, it is essential to understand not only the concepts and theories, but also the programs that make it possible, so it is recommended that you read the last chapter carefully.
Basically, each chapter is composed of Situation, Solution, Statistics, Application, Same application in different situations as mentioned above. However, please note that the chapters that do not require practice or that cover basic knowledge or explanations of statistical programs have slightly different structures.

Target audience

This book is written primarily for corporate managers who handle people data. We expect HR and marketing professionals to gain valuable insights and ideas for improving their work through this book.
Furthermore, I believe that graduate students, university officials, and researchers in the field of behavioral science will also be able to enhance their understanding of data analysis knowledge and program use through this book.


Writing team from the Educational Measurement and Statistics Laboratory at Korea University

The authors of this book specialize in educational measurement and statistics, but they are all deeply interested in applying measurement and statistical methods to corporate people data.
Professor Hong Se-hee, the corresponding author, studied industrial and organizational psychology for her master's degree and behavioral science quantitative psychology (quantitative psychology) for her doctorate.
Therefore, while I am very interested in research on pure quantitative development, I am also very interested in applying advanced quantitative methods to corporate data.
Based on this interest, the Educational Measurement and Statistics Laboratory at Korea University has been providing consulting services and conducting projects on data analysis to various companies.
Some examples of recent consultations or projects we have undertaken include:
?Development of a competency measurement tool and interpretation guide: We defined the competencies required by companies and developed a test tool to measure them.
An interpretation guide was also developed to be provided to each individual.
?Development of a fit index between competencies and jobs: An index was developed to show the degree to which a job is suitable based on an individual's competency pattern (e.g., high creativity, low relatedness).
We developed an algorithm that recommends the top five jobs that are well suited to a candidate based on a competency test, and developed a computer program that makes recommendations accordingly.

?Development of a quantitative model to predict job change: Most job change prediction models only predict the possibility (probability) of job change, but we developed a quantitative model that predicts the timing of job change (the point at which the probability of job change exceeds a certain level).
In addition to job changes, the occurrence and timing of specific events (e.g., promotions) can also be predicted in the same way.

?Convert scores from two different tests that measure the same thing so that they can be compared: If you've been using Test A, which measures a specific thing (e.g., job satisfaction), but for some reason you need to use Test B, you'll run into the problem of not being able to compare scores from the new test with scores from the old test.
However, the two test scores were equated to enable comparison through measurement theory.
?Developing a measurement method that encourages honest responses: The biggest problem in interpreting tests used by companies is that respondents do not answer honestly but rather answer in a desirable direction.
So when you score them, all the scores are high (e.g., all leadership scores are high).
We developed and scored a test that controlled for social desirability and faking.
?Deriving comparable scores by controlling for (taking into account) interviewer differences: Because interviewers vary in their rigor, evaluation scores from different interviewers are not comparable.
Interview scores were derived by controlling the interviewer's strictness.

?Measuring brain waves and determining status: Participated in the development of an algorithm that determines the status of a learner through bio-data measuring brain waves.
An interpretation guide for the results was developed.

It goes without saying that these various consulting and project implementation experiences were of great help in writing this book.


Consulting-related inquiries

For inquiries regarding advisory matters, please contact the corresponding author, Professor Hong Se-hee of the Education Measurement and Statistics Laboratory, Department of Education, Korea University, via email.
* seheehong@korea.ac.kr

Publication Plans for "People Analytics: Measurement Edition"

This book is titled “People Analytics: Data Analysis Edition,” and “People Analytics: Measurement Edition” will be published soon.
In the measurement section, we will cover techniques such as test development methods, reliability, validity, standard scoring, test validation, and item appropriateness analysis.
Measurement must come first for data analysis, but in most academic fields, objectively existing things are measured (e.g., sales volume, length of employment), so separate measurement techniques are not necessary.
However, variables such as job satisfaction, competence, and attitude are abstract concepts, and the measurement methods can vary, such as observation and self-reporting.
Another issue is how to achieve objective scoring.
It is not a simple matter to produce comparable scores, especially when there are different interviewers or multiple test scores.
Because education and psychology have always dealt with these variables, the field of measurement, namely educational measurement and psychometrics, has developed over a long period of time compared to other fields of study.
Here, we would like to introduce techniques developed in the field of educational measurement and statistics.

Finally, I would like to thank everyone who helped me publish this book.
All co-authors thank you for your hard work.
In particular, there was the hard work of Yang Jun-young, a doctoral student at Ohio State University.
Based on his experience working at LG Inhwa and LG Chemical before entering graduate school, Yang Jun-young devised realistic scenarios for each chapter, and he spared no effort in gathering all the co-authors' manuscripts and integrating and revising them right up until the very last moment before leaving for his studies abroad.
Express your gratitude.

Studying is just a painting.
Especially, in a field like this, which is not easy, only those who do not give up and study repeatedly can achieve it.
If you read and apply the materials to your own practice, you will gain a weapon with a high barrier to entry that others do not have. I hope you will continuously strive to acquire this weapon.

March 2021
Waiting for the end of Corona and the warm spring
In the laboratory of the Uncho First Education Center at Korea University
Hong Se-hee