Ludek's Blog About People Analytics
Divorce usually impacts both sides - but does that hold true for Brexit?
We can often come across analyses highlighting the negative effects of Brexit on the UK’s economic performance as measured by GDP. But what about the reverse perspective? How has Brexit impacted the economies of EU member countries?
"A new broom sweeps clean"
Does the saying above also apply to new managers and how they evaluate their direct reports?
Not all workplace relationships are created equal when it comes to retaining talent
A brief sneak peek into an interesting and potentially useful meta-analysis that explored the specific mechanisms through which two types of workplace networks are related to employee turnover.
ChatGPT as a new email writing coach?
Using CausalPy package to test the plausibility of the hypothesis that Stack Overflow may not be the only “victim” of ChatGPT and the GenAI likes.
How does your company stack up in the Big Nine Cultural Values?
A quick intro to the Culture 500 - a tool for assessing company culture using Glassdoor reviews.
Do you use XmR charts for People Analytics use cases?
Looking for tips from those who answered "Yes" to the question in the title of this post 🙂
Do people’s personalities vary across different jobs?
And if they do, are those differences big enough to be useful for career counseling and personal development?
Nonlinear relationships between personality traits and business outcomes seem to be the norm rather than the exception
... at least in my projects 😉
Simulating the "survivorship" effect in employee satisfaction data over time
How to align observations in organizational data with the results of one research on the relationship between tenure and employee satisfaction?
Motivation vs. Performance: What causes what?
An interesting research paper by Wang, Luan & Ma (2024) in Nature explores the causal relationship between work motivation and job performance using longitudinal data from 11 independent studies and meta-analytic structural equation modeling.
Why are introverts often described in terms of negative emotionality, and should they be?
If interested, here are a few quick thoughts of mine along with some light data exploration on the topic.
Are people during exit surveys more honest in their responses than in engagement surveys?
It is quite common to encounter the opinion that responses in exit surveys or interviews are more honest than those in regular engagement surveys, despite their anonymity and/or confidentiality, especially when it comes to more politically sensitive topics such as satisfaction with one's direct manager. It is usually argued that people no longer fear repercussions or negative consequences for speaking openly about their real experiences and opinions because they leave the organization and have less incentive to conceal critical feedback. Let's see if we can find any evidence of it in the data...
Unpacking surprises in the women’s 100m World Championships
How surprising was Sha’Carri Richardson’s victory in the women’s 100m race at the World Athletics Championships in Budapest 2023? Let's check it out with some data.
People may signal their exit intentions not only by their actions but also by their inactions
A brief reflection on one of the predictors of voluntary employee attrition.
Insights from the Team Assessment Survey results of C-suite teams
Results from the exploration of data on team effectiveness and efficiency across ~40 C-suite teams.
Kohonen's Self-Organizing Maps
A brief demonstration of an awesome tool for exploring multidimensional data, using personality and work-life satisfaction data.
Analyzing LinkedIn connections' jobs using LLMs and the BERTopic package
If you're curious about what your LinkedIn connections are up to, keep reading...
Does a stick work?
Probably each of us has experienced, or at least heard of, the practice of using job insecurity as a motivational tool at some point in our careers. A well-known example of this is stack-ranking performance reviews, where bonuses are given to top performers and those at the bottom are let go. While we may not personally like this method of motivation, it would be beneficial to have some data on its effectiveness.
Using network graph modeling to capture overarching thematic clusters in employee comments
A showcase on how to use network analysis to display the co-occurrence of topics in employee comments.
Can there be too much cognitive diversity in teams?
The answer might be "yes", at least in the case of team collective intelligence, defined as a general ability of a team to work together across a wide array of tasks.
Estimating body height using an inventory?
Why not?
What makes people more likely to comment on a question in an employee survey?
Is it satisfaction or dissatisfaction that drives comments? Or perhaps it’s the extremes on both ends of the satisfaction spectrum?
NetLogo: Don’t tell me, show me
An example of how to get a better understanding of various complex phenomena through simulation in NetLogo, a free programmable multi-agent modelling environment.
Feedback can be a gift or a curse...
...depending on where it directs your locus of attention.
If you are a leader, don't be afraid to be a nice one
Do nice guys really finish last?
Does agile pay off?
During my career, I have been part of several teams that have organized their work using some of the existing agile methodologies. For the most part, agile principles made good sense to me and subjectively seemed to work, so I had no reason to question their supposed benefits.
A showdown between the Big Five, Enneagram, MBTI, and astrology
Who you got your money on? 😉
Does your team belong among “light bulbs” or “wines”?
Sharing one learning from the awesome book Probably Overthinking It by Allen B. Downey.
Biases introduced by video backgrounds during video interviews
Every tool used for employee selection introduces its own potential biases into the process. This holds true also for video interviews, which have become increasingly popular among recruiters in recent years.
Creating new candidate topic labels on the fly during topic analysis with GenAI
Description of a simple hack to simulate the work of a qualitative researcher classifying comments from respondents using GenAI.
Fostering psychological safety in the workplace
Can a simple intervention such as an email campaign help with fostering psychological safety in the workplace?
Fighting with faking in personality assessments
A brief summary of the main results of a meta-analysis comparing psychometric characteristics of forced-choice and single-stimulus personality assessments in relation to faking.
How should teams be designed to be creative and innovative?
A brief overview of the results of a meta-analysis on the relationship between team design and team creativity and innovation.
Strength-based development and power-law vs. normal distribution of performance
Does a power-law distribution of performance, as opposed to a normal distribution, support the concept of strength-based development?
Do high-low range goals aid in maintaining motivation over time?
Is there enough evidence to bet on this technique and give it a try?
Does GenAI make me a better (more rational) thinker?
A short reflection on one of the impacts of using GenAI on my work habits.
Reflective and intuitive thinking styles
A new construct on the block for employee selection and development?
Link between the Big Five personality traits and earnings
Check how your personality supports your earnings.
How to support the adoption of Evidence-Based Management with a specialized GPT bot
Let's create your own consultant to help you apply the principles of Evidence-Based Management.
Rebuilding an employee survey with the help of NLP tools
Using LLM and text embeddings to assist in implementing new constructs into the existing employee survey.
Evidence on the presence of gender bias in selection settings
Interesting results from a pre-registered meta-analysis of 44 years of field experiments on gender gaps in hiring decisions.
Surprising finding on the impact of job demands and control on workers’ well-being
I just came across an interesting and surprising result from a Bayesian meta-analysis on the effect of the interaction between job demands and job control on worker well-being.
Personas based on ML local interpretation algorithms
A demonstration of one method useful for sharing insights from fitted ML models.
Sentiment analysis of employee survey comments using zero-shot classification
An attempt to validate a zero-shot sentiment classification.
Using NLP & LLM to combat 'tip-of-the-tongue' moments during onboarding
How to make onboarding experience a little bit smoother with the help of NLP and LLM.
Exploration vs. Exploitation trade-off in our calendars
As I was going through my calendar recently to check who I had already met during my onboarding at Sanofi, I realized that one way to look at the calendar is through the lens of the Exploration vs. Exploitation trade-off. What lessons can we take from this?
Harnessing Bayesian analysis for business process simulation
A demonstration of how the outputs of Bayesian analysis can be used to simulate business processes while preserving inherent uncertainties.
A plausible model of data-generating process eats ML algorithms for breakfast
An illustration of one of the lessons I took away from studying the use of meta-learners for causal inference.
Embedding R (or Python) ML models in Power BI dashboards
In my new job, we currently rely a lot on Power BI when presenting people-related insights to our stakeholders. Since I know Power BI quite superficially and we also want to share insights from more complex analyses with our stakeholders, I spent part of the weekend studying how to incorporate ML models created in R or Python into Power BI dashboards. I put my learnings in this blog post. It's definitely not rocket science, but it may still shorten the learning path for some of you who are in a similar situation.
Vocational interests don't seem so uninteresting after all
Quite surprising (at least to me) findings on the validity of vocational interests for predicting a range of important work outcomes.
Induced centralities
A post about useful complement to common ONA centrality measures.
Searching & querying AIHR blog posts on People Analytics topics
I'm sharing a by-product of my learning about vector database search that may be useful to some of you who want to learn something new about People Analytics.
Using Bayesian shrinkage in reporting employee turnover
When you report turnover rates by team, do you take into account the size of individual teams, or do you take the turnover rate numbers as they are?
A bet on a new job
Sharing a by-product of my search for a new full-time job.
Themes in Deloitte's Global Human Capital Trends between 2011 and 2023
Reading the latest release of Deloitte Global HC Trends made me wonder what common themes this regular series has been covering throughout its 12 years long history.
Team-level predictors of innovation at work
Team processes seem to beat team composition and structure when it comes to innovation at work.
Latent Class Analysis of responses from employee surveys
How listening to a podcast about conspiracies and disinformation inspired me to try out a "new" statistical tool popular among sociologists.
Another positive effect of mindfulness meditation on the horizon?
Is it possibile to improve the objectivity of decision making through mindfulness meditation?
Use of Bayesian networks in people analytics?
Bayesian networks seem to have some interesting properties that could make them useful for various people analytics use cases, but for some reason this is not the case.
Visualizing shifts in validity estimates for selection procedures
Let's take a slopegraph perspective to assess changes in estimates of the validity of selection procedures.
Psychometric network analysis & employee survey data
A demonstration of how psychometric network analysis can be used to gain insights into employee survey data.
Employee commitment over time & regression to the mean
A nice illustration of the regression to the mean phenomenon in the space of people analytics.
Improving a company career site with tools from OpenAI
How my own experience of exploring new job opportunities gave me the idea of how the company's career site could be easily improved using OpenAI's tools.
GPT-4's performance in the knowledge test of evidence-based HRM practices
How did GPT-4 perform in the knowledge test of evidence-based HRM practices? Let's check it out.
Employee feedback analysis using tools from OpenAI
How to use GPT and embeddings from OpenAI for identifying topics and related sentiments in employee feedback.
When flawed statistical & causal reasoning leads to a valid conclusion anyway
Comparison of Glassdoor ratings from current and former employees.
In need of multilevel correlations?
A post about a great R package to reach for when you need to calculate correlations on nested data.
Consequences of time management in the workplace
Some interesting insights from a meta-analytic review of the consequences of time management behaviors in the workplace.
Interpretable machine learning with modelStudio
There's a new kid on the block in the R ecosystem that can help analysts understand the behavior of their ML models.
Ask your personality using GPT
Can Generative AI like GPT meaningfully interpret personality profiles?
Effectiveness of interventions for encreasing employee engagement
What evidence do we have for the effectiveness of interventions for increasing employee engagement? Let's check it out.
Where do managers put on their rose-tinted glasses the most?
In which areas are managers and leaders prone to overconfidence, and how can this overconfidence potentially impact team functioning? Let's check some data to address this question.
Estimating the impact of a new business app by piloting & method of difference-in-differences
What is the benefit of using the difference-in-differences method in combination with piloting a new business app, and how can this help estimate the app's effectiveness on key outcomes like time spent with prospects or closed deals?
Are there meeting “yes-men”?
One of our clients was struggling with meeting overload and wanted to know if the people who attend too many meetings are the kind of "yes-men" who just can't say no to meeting invites. You know the type - always saying "yes" and never protecting their precious time. What did they find?
Impact of employee satisfaction at work on a company's bottom line
While there is evidence supporting the connection between employee satisfaction and a company's bottom line, it's essential to determine whether higher satisfaction directly causes better performance. Is there some evidence for that? Let's check it out.
Excel + Python = Word Document
Using combination of Excel and Python for semi-automatic Word document generation.
Employee outcomes & employees' job attitudes
What employee outcomes are predicted by what employees' job attitudes? Let's check it out.
Effectiveness of micro-breaks at work
Quite satisfying news from a meta-analysis on the efficacy of micro-breaks for increasing well-being and performance in the workplace.
Not-so-hidden cost of working in an office
Just a few data-backed thoughts on why many of us may often feel more distracted when working in an office.
Using ChatGPT to summarize and explore employee feedback?
What's the potential use of tools like ChatGPT in analyzing open-ended feedback from employee engagement and satisfaction surveys? Let's take a look at the result of my little experiment in this area.
Where to look first when considering meeting reset?
Let's briefly discuss the potential benefits of focusing on optimizing large recurring meetings to save time in the workplace.
HRM value chain and structural equation modeling - Moneyball case
What's the link between the HRM value chain and structural equation modeling? Let’s check it out.
Org chart and collaboration
How to effectively combine information about the formal organizational structure of a company and the actual collaborative activities of its employees?
Span of control and collaboration data
How can collaboration data be used to determine the "optimal" scope of control?
Attaching numbers to best practices for instant messaging
Slack and other instant messaging platforms can be both a blessing and a curse. Can we attach numbers to some of the recommendations on how to use them effectively? Let's take a look.
Always messaging
Let's take a look at two concepts from computer science that can be used in the workplace to improve people's focus and productivity, and expose two methods for measuring their related behaviors when collaborating on instant messaging platforms.
Warning system for overloaded employees
What tools and/or signals can we use to identify employees at increased risk of overload? Let's take a look at some of the options we have in this regard.
Evaluation of the results of the evidence-based HRM knowledge test
Have we made any progress in knowledge of evidence-based HRM practices in the last 20 years? Apparently not. But let's look at the details.
How to improve effectiveness of meetings?
What suggestions do people have for improving the effectiveness of meetings? Let's check it out.
Why are meetings in/effective?
If you are a regular organiser or attendee of meetings, you may be interested in what people think about the reasons why the meetings they attend are in/effective, as this can give you a better chance of contributing to making your meetings more effective and meaningful for you and others.
Signals of a good manager
Do you think it's possible to find signals in collaboration (meta)data that someone is a good manager? Let's give it some thought.
Evidence-based HRM knowledge test
Interested in testing your knowledge of evidence-based HRM practices? If so, click and get started.
Can you really unplug?
With the Christmas holidays approaching, the following question is more relevant than ever, with the exception of the summer vacations: Can we really disconnect from work during the vacations? And what can collaboration data tell us about this?
One does not simply do a business without getting lost
Breaking down one weekend association.
How to quickly navigate dashboard users to what they need to know?
Let's take a look at some tips and tricks to make dasboards more useful for their users.
Eisenhower matrix for meetings
Meet the Eisenhower matrix for meetings ;)
Some resources on staff retention and downsizing
Probably due to the current situation in the talent market, where many companies are laying people off and at the same time are worried about losing their key employees, a few people have contacted me in recent weeks asking for some tips on evidence-based approaches to dealing with retention and downsizing.
Evidence-based interventions that help reduce the gender pay gap
Pay inequality between men and women is not only an ethical and legal issue for companies, but also a marketing issue - it can have a negative impact on their "employer brand" and attractiveness as an employer. This means that if companies want to attract and retain talented employees, they must be able to ensure that they treat men and women equally in this respect. Let's look at what the existing evidence tells us about what might help us with this.
It's perfectly normal not to be normal
And it definitely applies to the shape of the distribution of many HR metrics. Let's look at this in a little more detail.
A visual introduction to Bayesian belief updating
Teacher: "Bayesian belief updating involves combining existing or prior beliefs with an assessment of the strength of new evidence." Student: "And could I please see this in action?"
Police cadet evaluation dataset
A "new" real-world dataset useful for training in people analytics.
Divide and... understand
Finding the breakpoint when people start to score significantly higher/lower on a given criterion - the use case for the Conditional Inference Tree algorithm.
Timeboxing. Does it really work?
Checking with real-world collaboration data whether timeboxing has a protective function in terms of time available for focused work.
Multilevel modeling in people analytics
Don't chase (statistical) ghosts and use multilevel models instead!
Collaboration and personality
Personality is not fate, at least when it comes to the level of engagement in corporate communication and collaboration.
Hot spots of collaboration overload and collaboration bottlenecks and how to find them
One of the most useful insights that can be gleaned from collaboration data is where hot spots of potential collaboration overload and/or collaboration bottlenecks may exist in a company. Such insight can be especially valuable these days, when many companies are trying to fight the upcoming economic downturn by achieving more with less.
Are back-to-back meetings for good or bad?
A short post about the practice of back-to-back meetings and how to determine when it's for bad and when it's rather for good.
The impact of the COVID pandemic on the popularity of people analytics
Many people analytics professionals think that after the COVID pandemic, organizations are more willing to listen to their insights and recommendations. Can we find any empirical support for their hunch? Let's check it out with data provided by Google Trends and segmented regression analysis of interrupted time series.
People Analytics Challenge from Orgnostic: Plan for high growth
A brief summary of my participation in Orgnostic's People Analytics Challenge.
Visual statistical inference
Visual statistical inference represents a valid alternative to standard statistical inference, and as a by-product it also helps with building intuition about the difference between signal and noise. Give it a try.
Standard vs. trend predictors
When modeling a phenomenon, one usually can't get by with just raw data but must use one's domain knowledge to select and transform the most relevant variables from raw data to be able to successfully grasp regularities in the domain of one's interest. Let's look at one simple example of such feature engineering from the domain of collaboration analytics.
Fighting meeting overload
One of the most effective ways to fight meeting overload is to better plan meetings in terms of the time we spend in them. Let's look at how data can tell us how much room for improvement we have in this area.
How do we perceive probability words?
Have you ever wondered exactly how much chance of success people give a project when they say they believe in it? If so, then you may find this post useful, as it attempts to answer that question at least in part with data.
Hofstede's theory of cultural dimensions
Cultural diversity brings both positive effects and some challenges. To deal with the latter, it is useful to have some kind of map to help people better navigate the cultural specificities of people from different societies. Hofstede's theory of cultural dimensions is useful for such a purpose. Let's check how dis/similar countries are on these cultural dimensions with a simple app that could help us better understand, manage and appreciate cultural differences a little better.
Makers' schedule and managers' schedule in collaboration data
Many of us have probably already heard of Paul Graham's two types of schedules - one that meets the needs of makers and one that meets the needs of managers. But can these two types of schedules be found in any real collaborative data? Let's find out.
R Shiny app for LinkedIn connections analysis
An introduction of a simple R Shiny application for analysing LinkedIn connections.
Overview of predictors of voluntary employee turnover
An introduction of a simple R Shiny application to facilitate extraction and digestion of information from meta-analysis of predictors of voluntary employee turnover.
Firemní audit rozdílu mezi platy mužů a žen
Platová nerovnost mezi muži a ženami není pro firmy jen záležitostí etickou a právní, ale také marketingovou - může mít totiž negativní dopad na jejich "employer brand" a atraktivitu coby zaměstnavatele. To znamená, že pokud firmy chtějí přilákat a také si udržet talentované zaměstnance, musí být schopny zajistit, že se u nich s muži a ženami bude v tomto ohledu zacházet stejně. Prvním krokem k tomu je zjistit, jak velký je rozdíl mezi platy mužů a žen ve firmě a do jaké míry ho lze vysvětlit jinými faktory než je samotné pohlaví zaměstnance. V tomto článku demonstruji, jak takovou analýzu provést s pomocí analytického nástroje R a dat, která má většina firem běžně k dispozici. Stručně se zmiňuji rovněž o tom, jaké mohou být případné další kroky a doporučení vyplývající z výsledků provedné analýzy.
Modeling impact of the COVID-19 pandemic on people’s interest in work-life balance and well-being
Illustration of Bayesian segmented regression analysis of interrupted time series data with a testing hypothesis about the impact of the COVID-19 pandemic on increase in people's search interest in work-life balance and well-being.
HR analytika a odchodovost zaměstnanců
Které faktory přispívají k odchodovosti zaměstnanců a u kterých konkrétních zaměstnanců je zvýšené riziko, že firmu během několika příštích měsíců opustí? Na tyto otázky se čím dál tím více firem snaží odpovědět pomocí analýzy dat o svých vlastních zaměstnancích. V tomto článku se prostřednictvím analytického nástroje R a vizualizačního nástroje Shiny podíváme, jak může být tento druh HR analytického projektu pro firmy užitečný.
Moneyball v HR
Přes popularitu tématu HR analytiky mezi HR profesionály je stále relativně málo společností, které HR analytiku reálně a systematicky využívají. Jednou z možných příčin je to, že tradiční HR mnohdy postrádá analytický mindset a některé z kompetencí, které jsou klíčové pro úspěšnou realizaci HR analytických projektů. V takové situaci může být užitečné podívat se ve větším detailu na celkovou logiku i na konkrétní analytické kroky nějakého úspěšného příkladu využití HR analytiky k optimalizaci některého z HR procesů s pozitivním dopadem na obchodní výsledky společnosti. V tomto článku se tímto způsobem podíváme na známý příběh oaklandského baseballového týmu "Áček", jehož management poměrně radikálně - a podle všeho i úspěšně - přehodnotil svůj dosavadní přístup k výběru nových hráčů na základě výstupů statistické analýzy sabermetrických dat o herním chování hráčů. Využijeme při tom volně dostupný statistický software R a veřejně dostupnou databázi historických údajů o výsledcích v americké baseballové lize.