ASA Connect

 View Only
Back to eGroups
Expand all | Collapse all

What are your "real world" experiences at your jobs? (Guidance for students)

  • 1.  What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-11-2020 17:05
    Hello community!

    I've reached out in the past for your help (do you remember the conversation about graphs?) and I'm thankful for your advice!

    Today's "ask" is about "real world" experiences. Following up on this article from Faculty Focus about teaching students for the real world, please tell me about your job or past jobs that college didn't (or did!) prepare your for. 

    For example, I teach online, community college introductory statistics and our examples from the textbook are tidy little examples of data confidence intervals or hypothesis tests. Sure, assignments have due dates, but I want to teach my students more than just "turn in your work...and I'll probably accept it late, too."  My own "real world" consists of teaching statistics so I feel like a fraud trying to tell my students about the "real world" pressures, challenges, and experiences that non-teachers face. 

    An English instructor told me today she was asked to rewrite a grant in a few hours, that just landed on her desk. Now that's a challenge! I can see myself making a (semi?) surprise activity that's due the next day, which challenges students to think quickly, be resourceful, and think about their own time management.

    If I were to to tell my students "In the "real world" you'll need to know how to __________, so I want to prepare you for that," how would you fill in the blank? I'd love to hear anecdotes or advice, if you'd like to share them. :)

    Thank you! :D

    Jennifer

    ------------------------------
    Jennifer Ward
    Adjunct Instructor, Clark College
    ------------------------------


  • 2.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-12-2020 07:38
    Hi Jennifer,

    I'll be the first penguin in the water....

    "In the real world, you'll need to know to ____plot/display data____."

    In many instances (even for data-focused, not-unsophisticated companies) there is a heavy need to simply display the data available. 
    Even for someone who will eventually need to analyze that data, plotting it is the first step to understanding.

    People outside of the data analysis role will typically not have opinions (that are both strong and informed) on how the data should be analyzed. But you should see their eyes light up when they see their data (plotted) for the first time. This may seem trivial to people focused on providing sophisticated analysis...but it has a huge value to the data owners whom we aim to serve.

    A common follow up to this is asking the data analyst "can you make more plots like this?" or "can we have these plots updated as new data comes in?"

    So that's my real-world observation. My great analytical passion is using Bayesian non-parametric modelling of timeseries for novelty detection....but day-to-day, my value comes from getting into the non-analyst's head and displaying data for non-analysts to understand. (The backend operations to robustly display data can become fairly statistical in it's own right...depending on how dynamic the data is. However, the data user is unlikely to consider these elements, they'll just wonder why the graph is ugly/ill-proportioned/hiding data.)



    ------------------------------
    Glen Wright Colopy
    DPhil Oxon
    Data Scientist at Cenduit LLC, Durham, NC
    ------------------------------

    Original Message


  • 3.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-13-2020 09:14
    Learn the business in which you are working. This will help you understand when you have bad data and analytical results that just don't make sense.

    I worked in numerous areas over my career (gov't, marketing, finance, insurance) and in each one I needed to learn from Subject Matter Experts what made sense and what was nonsense.

    Knowing what looks right will not only help build better models and predictive results but will also help tell stories about the analysis.

    DON'T just jump right in and tell your customer how things look without first running the results past people who know the business.

    ------------------------------
    Michael Mout
    MIKS
    ------------------------------

    Original Message


  • 4.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 14:02
    I've found in many instances, the results that don't make sense are the ones you want to focus on. 

    I gave a presentation to a VP of QC. The only important factor was manufacturing date. The dates that were important were days when the factory was closed/shut down. The results didn't make nay sense to us or our liaison. Turns out, that was the important thing.  The VP instantly knew what was going on. No one else did. 

    If you go back and look at some of the articles where the authors used Design of Experiment methods, you find that a lot of the time, the results are better than the standard method and tend to contradict what is "known" about the system. In fact, George Box's paper on the first Central Composite Design shows that the chemists and chemical engineers that had been working with the chemical system for years had it all wrong! 

    Just because it doesn't make sense to experts doesn't mean its wrong. Experts have different points of view and different areas of expertise. An example, if you look at the concentration of a certain isotope in Martian soil samples (if I remember correctly) are so high that either Mars was bombed with multiple (hundreds to thousands of) Empire State Building sized nuclear bombs millions of years ago OR there is a new atmospheric phenomenon. 

    When analyzing data, I would stick with one of George Box's quotes, "Theory Guides: Data Decides."

    ------------------------------
    Andrew Ekstrom

    Statistician, Chemist, HPC Abuser;-)
    ------------------------------

    Original Message


  • 5.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-13-2020 09:58
    I'd like to second Glen's point and expand a bit given my current and past employers.

    While meetings/presentations rarely get pushed back, data always comes in late. This requires a very short turn-around time for being able to summarize the data in meetings. One of the best ways to do this is through plots/graphics where it's much easier to walk the audience through the results. 

    Therefore, I'd recommend maybe doing an assignment where you give the students a small dataset and ask them to prepare a 5-minute presentation on the data. This would allow them to plot the data and maybe do some simple t-test comparisons to look for differences between two groups, like the results of a particular question by gender. This could also drive the point that two groups NOT being statistically different is also a legitimate finding in analyses.

    ------------------------------
    Martha McRoy
    International Research Methodologist
    Pew Research Center
    ------------------------------

    Original Message


  • 6.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-14-2020 01:46
    Funny enough, Glen, the mascot here at Clark College is the Penguin! Oswald the Penguin, to be precise. :D

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 7.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-13-2020 08:17
    Hi Jennifer,

    What a great idea.  Your students are lucky to have you.  I saw the response from Glen and I couldn't agree more.  But here is another.​ 

    "In the "real world" you'll need to know how to __________, TELL STORIES.

    You need to logically explain the conclusions.  Why do the conclusions and proposed recommendations make sense.  This is best done through stories that make sense to your customers.  Would you select an 80% free throw shooter (based on 100 attempts) or a person that made 1 out of 1 free throws for your basketball team (variability around an estimate).  Would it make a difference if I told you that the 80% free throw shooter was a professional and the other person was randomly selected off of the street, or the reverse scenario (prior knowledge).

    Would you be comfortable NOT rejecting the null hypothesis with a p-value of 0.06 if the Null Hypothesis was "This person is not guilty of shoplifting"?  What if the null hypothesis was "This sleep aid will give you brain damage".

    In order to prepare and explain the best recommendations, the problem and its ramifications need to be understood.


    ------------------------------
    Philip R. Scinto
    Senior Fellow
    The Lubrizol Corporation
    ------------------------------

    Original Message


  • 8.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-19-2020 18:25
    Hi Philip, 

    I REALLY like this simple statement of telling stories. I even used that line to help write a question on the final exam for this week.

    Next term I'm going to use it as an overarching theme of the term, and weave it into all the work I make students do. How fun is it to think about statisticians as storytellers? I think I can allay the fears that many community college students have about a math class by thinking of themselves as storytellers. :)

    Jennifer

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 9.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-13-2020 08:38
    In the "real world" you'll need to know how to MAINTAIN A PROFESSIONAL DEMEANOR and FIT INTO A COMPANY'S CULTURE.  

    You can be perfect at your job, but to open up opportunities for yourself and safeguard yourself from personality conflicts, you need these two skills specifically. We statisticians have a reputation of not being very good at them.

    ------------------------------
    Sheila Braun
    ------------------------------

    Original Message


  • 10.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-14-2020 01:45
    Thank you, everyone who has replied and emailed me your suggestions for advice for students! I plan to compile a webpage for my class with your wisdom.

    Amid the flurry of schools emails about Covid-19, each response has been a breath of fresh air! :)

    Jennifer


    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 11.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-16-2020 09:56
    "In the "real world" you'll need to know how to ___WORK WITH IDIOTS_______, so I want to prepare you for that,"

    I've run into people who have no idea what they're doing or what they want.  Some even have Ph.D.s in the subjects they don't know.  (Fortunately, this is the minority.)

    I and other consultants often receive calls for consulting from people who don't know what they want.  The consultation is really a therapy session.  Usually, you'll be able to reach an understanding.  I've been able to point a few people in the right direction.  But, then, there was the character who wanted to develop an automated statistical package to do statistics without a statistician.

    Many people don't know what they don't know.  I once was at an office meeting in which we welcomed a new President.  All of the VPs introduced themselves as computer experts, even though many barely knew where the switch was.  The office manager was the true expert because she knew what she didn't know.  I've seen more than a few papers in which doctors played statisticians and published nearly useless papers that other doctors proceeded to use.

    On the job, I've had to deal with people who didn't know their own fields, so it was extremely difficult to     communicate with them about their own subjects.

    ------------------------------
    Chuck Coleman
    ------------------------------

    Original Message


  • 12.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-17-2020 10:50
    I have to raise a strong objection against Chuck Coleman's comments about "working with idiots." 

    The people I have worked with, almost without exception, are very smart. Some of them may know Statistics really well and some do not, but just about every one of them knows their own subject area better than I do. That doesn't mean that they don't have blind spots, and I am very quick to raise scientific concerns that I might have even if they are outside the realm of Statistics. But I don't raise a scientific issue without careful thought, and I listen closely when a scientist offers a defense of what might seem to me to be a misguided scientific approach.

    The reason that Statisticians are marginalized is because too many of us act like we know more about the problem than anyone else.If you want to be a good statistical consultant, you need to go into a consulting session with the expectation that you will learn as much if not more about the science as they will learn about Statistics. If you want to be a lousy statistical consultant, go into a consulting session thinking "here's another idiot in need of my wisdom."

    ------------------------------
    Stephen Simon, blog.pmean.com
    Independent Statistical Consultant
    P. Mean Consulting
    ------------------------------

    Original Message


  • 13.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-17-2020 11:12
    As I said, in any statistical work, the first thing you need to do is learn as much as you can about the business and the best person to get that information i a subject matter  expert. Make that your first job.

    ------------------------------
    Michael Mout
    MIKS
    ------------------------------

    Original Message


  • 14.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-18-2020 13:22
    My advice would be to learn as much as you can about the data generating process. Go and see where the action happens if possible, even if it is just someone clicking things. Pay attention, with caution, to anecdotal comments (normally in this range, never seen value more than x). Sure, you can't replace subject matter expertise, but this process at least gives the ability to perform a first pass sanity check. Additionally, doing EDA (which should always be done on a new data set) you could spot the outliers (not in a statistical sense) or those points that don't fit the expectations. You could add this into an assignment prompt by saying the machine detects from this range (e.g. 0, 100) and have a few points that have values of -1 or 110 which could be errors on a sensor or something like this. Maybe you see that the most common reason code for a machine stop or diagnosis or whatever is the first item on the list (so what is real vs what is just someone selecting the first pop-up). Or perhaps include a spike in the data that relates to the super bowl or an election, just to get people thinking about external forces.

    Another thing is that you need to treat everything you do, each analysis, firstly that this won't be the last time you do it (i.e. someone will ask you to repeat this work for the same scenario or a different scenario/data/etc) and secondly that eventually you will hand off the work to someone else. Let that shape how you do your work, document your code, assumptions, etc. It would be interesting to have a group assignment where you have to do a project, then hand it off to another person with a slightly different data set and try to get the same results. But that's the real world.

    In the real world, you sometimes have to give a good enough answer and sometimes that means I don't know. We don't have enough data to conclude there is a difference.

    Finally, the "always be integrating over your loss function" concept. What happens if we are wrong? If I say there is a difference and there isn't, how much did I cost the company or how many unnecessary test did I run? I think doing statistics in light of the real life loss function or the "cost" of error is a really important concept. Hey, if I'm wrong by 50%, we still do no harm so let's do it vs Hey this effect is really small and significant, but we run a huge cost of xyz...



    ------------------------------
    Michael DeWitt
    michaeldewittjr.com
    ------------------------------

    Original Message


  • 15.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-19-2020 18:21
    Thank you, Chuck, Michael^2, and Stephen for your perspectives! I appreciate that you each can bring something to the table.

    I've already started summarizing the big picture ideas from this thread and will be excited to share it with my students...whenever the term starts. :)

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 16.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-17-2020 08:17

    I find that even in places and among people that are doing advanced statistics, the most important work involves very basic things that involve understanding the implications of the problem they are looking at.

    I once worked for a client where for many years, management had been told that resource utilization for projects started a quarter ago was less than for projects started in the same quarter a year ago, thus showing improvement in efficient (more recent projects have less time to use resources.) In another situation, management was based on average resources per project calculated at project end, but a handful of extreme cases would completely throw the average and suddenly make the subdivision look bad by the formula. (An average isn't a reliable estimate for an extreme value distribution). 


    Medical journals regularly publish clinical trials where the reported median occurred at the first assessment in both arms, and the medians are reported as being the same with high reliability due to narrow confidence intervals. (In this case, the actual medians occurred before the first assessment and went unobserved, and the tight-looking confidence intervals reflect only scheduling variation around the target assessment date and not variation in the thing assessed.) I have seen a study report that confidentially reported no change in outcomes since the last report, when what happened was that the outcomes stopped being assessed after the previous report.

    I could go on.


    To be effective in a real-world setting, the most value often comes through asking basic, simple questions about what is really going on, how the data is being collected, what reality it actually corresponded to, what it actually means.

    This is often far more important than doing a sophisticated analysis. I have seen very sophisticated analyses that reached misleading conclusions by ignoring basics. 


    Of course, learning how to mention things like this with tact to people with big egos, high positions, and advanced degrees is also an art, and can be even more important when it comes to keeping a job. 



    ------------------------------
    Jonathan Siegel
    Director Clinical Statistics
    ------------------------------

    Original Message


  • 17.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 13:15
    Thank you, Jonathan! You make a good case for why students should have a strong understanding of the fundamentals as they leave my class. Having skills to ask questions tactfully, when one doesn't understand what's going on, is also worthy.

    Jennifer

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 18.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-17-2020 15:32
    In the real world you need to know how to say "no". 

    In the profession, we will deal with a lot of people that will:

    1) P-hack
    2) Insist on inappropriate analyses
    3) not know what is possible
    4) not know about the "rules" of analysis

    P-hackers are everywhere. Insert your own stories.

    As a chemistry and physics student, we "learned" about all the ways to "properly" design an experiment. Should one compare those methods to what is taught in a textbook written by a statistician or industrial engineer, one will find most of what is taught by scientists is simply wrong. Often so wrong that they are the opposite of what you SHOULD do. 

    How do you tell someone that using multiple t-tests is a bad idea. How do you explain why you need to use corrections? How do you tell someone that you can and should change multiple factors and levels simultaneously during an experiment? How do you explain why categorizing continuous variables can be a really bad idea? 

    Being able to say, "No. That is a wrong approach. Let's try ____." Is very important. 

    5) tell an expert statistician their assumptions about the data are wrong

    As a graduate student, I wanted to write an article about the use of Mixture Designs for toxicology tests. I asked the head of the statistical consulting group about this. They had designed dozens of toxicology experiments. They had never heard of anyone testing more than 2 chems at a time. They assumed no one did more than one at a time because you cant and dont need to because chemical interactions where rare. When I asked about the warnings you see on medications, their interest changed.

    ------------------------------
    Andrew Ekstrom

    Statistician, Chemist, HPC Abuser;-)
    ------------------------------

    Original Message


  • 19.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-17-2020 17:26
    Learn how to use XKCD comics to great effect. For instance, the jelly bean comic is a great way to teach about multiple t tests. I'm not sure I could teach stats effectively without XKCD comics. 

    XKCD Jelly Beans - Significant

    ------------------------------
    Sheila Braun
    ------------------------------

    Original Message


  • 20.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 13:40
    Hi Sheila!

    In fact, I start every module with one of the XKCD comics! :D

    You have sparked an idea that I could use the comics as a topic of discussion in small groups, instead of "just something" that students read at the start of a module (and probably ignore?).

    Jennifer

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 21.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-18-2020 11:21
    ​"In the "real world" you'll need to know how to clean and monitor data."

    This will be 70% of your job. There are some tools to clean up values based on appropriate range, appropriate string characters, etc. but most errors will be in record linkage. Lots of data is still entered by people who will inevitably mistype or misread someone's notes (the notes will also have errors), or is affected by software compatibility and updates. As many databases are in constant live update, cleaning will be ongoing.

    Another 20% of your job will be building a data file capable of yielding meaningful results. What distinguishes a duplicate record from a follow up and what is an unbiased way to choose?


    ------------------------------
    Shalese Fitzgerald
    University of Arkansas for Medical Sciences
    ------------------------------

    Original Message


  • 22.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-18-2020 11:30
    And part of that job will be reviewing basic stats of the variables, looking for unusual outliers or odd tendency (both central and disperse). The key being to try to find the bad data. All of which requires knowledge of the data and environment.


    ------------------------------
    Michael Mout
    MIKS
    ------------------------------

    Original Message


  • 23.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-18-2020 11:49
    Shalese's answer is definitely worth a second read. Perhaps, Jennifer, when you relay these answers to your students, you should have then read through twice - the first time to grasp the breadth, and the second time to catch the repeated themes.

    Ex. "clean and monitor data....many databases are in constant live update, cleaning will be ongoing."
    This is definitely a big thing and Shalese is giving a very useful heads-up by including the "monitor", because it's a hefty task.

    Typically when you clean a data set the first time you will see a combination of (i) one-off errors, and (ii) systematic errors in the data.
    [Please include comments if there's a substantial category that's I'm missing, ex. "(iii) missing data".]

    A broad generalization may be that human-generated records may tend towards 1-off errors, whereas computer-generated data may tend towards consistent/persistent errors.

    After cleaning a data set once, you'll realize that the next set of data coming in will likely be plagued with similar issues. You may even wish to proactively prevent future errors making it on to the record by engaging with those generating the data. This is where the monitoring comes in. If you are working in a domain that values novelty detection, then congratulations: data cleaning and monitoring will be even more deeply embedded in your analysis.

    While this might seem a bit vague to new students who haven't worked on a long-term data problem, it's a real thing for practitioners.
    So perhaps a good takeaway is "don't get discouraged if you spend a lot of time cleaning, organizing, recleaning, monitoring, and cleaning your data."



    ------------------------------
    Glen Wright Colopy
    DPhil Oxon
    Data Scientist at Cenduit LLC, Durham, NC
    ------------------------------

    Original Message


  • 24.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-23-2020 14:14
    What about define "cleaning of the data"?

    ------------------------------
    Andrew Ekstrom

    Statistician, Chemist, HPC Abuser;-)
    ------------------------------

    Original Message


  • 25.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-24-2020 11:43
    That's an interesting question and, as Glen said, one that quickly takes on enormous breadth and depth.
    As a general guideline, we might say data cleaning is the identification and decisions surrounding data that is inconsistent with possible values.
    When I write code for data cleaning, the majority of it goes toward defining the space of possible values.

    ------------------------------
    Shalese Fitzgerald
    University of Arkansas for Medical Sciences
    ------------------------------

    Original Message


  • 26.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-24-2020 13:28
    One of the great ways I learned to clean data, especially in model development regimes, was from FICO. The idea was to not only look at the basic stats of all the variables, but also review the bi-variate relationships between a subset of the independent variables and the dependent variable (assuming one dependent variable). In that detailed review one would look at whether the relationship was linear, whether it made sense (was in the correct direction) and if there were any weird outliers.

    This allowed the modeler to eliminate any nonsensical relationships, or, if they were strong enough, to try to find a way to explain those relationships. Same for non-linear relationships, eliminate or find a say to explain and transform.

    It is a time consuming process, but results in better models that make sense and endure over time.

    Over the years I taught this process to modelers who worked with me.

    ------------------------------
    Michael Mout
    MIKS
    ------------------------------

    Original Message


  • 27.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 13:38

    Hi Glen, Andrew, Shalese, and Michael, 

    After taking part in the StatPREP Seattle Hub the last two years I do now include an activity in all my classes about cleaning data. Danny Kaplan really impressed it upon us that data in stats books are not an authentic experience.

    I have included an activity for students to compute an average number of pets that students have, where one of the responses is "a lot of fish". After reading y'all's responses I think I can give them a messier data set, and ask them how long it took them to get an answer. Not only will students be able to discuss the struggles of computing an average, be able to compare their responses, but also see *just how long* they spent getting ready to compute the average. Oh man, this is going to be exciting!

    Jennifer



    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


  • 28.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 14:06
    Back when I taught Intro to Stats I would regularly have the students design their own experiments and collect their own data and then calculate the specific exercise on which we were working (Mean, Median, T-test, ...). Part of the exercise was to write up the result and describe issues with the data.

    Worked fine for small classes (<15) but a bit of a pain to grade with larger classes.

    ------------------------------
    Michael Mout
    MIKS
    ------------------------------

    Original Message


  • 29.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-25-2020 14:26
    I have my Mgt Sci students generate data, then find averages, percentiles, etc. Then comment about the data. (I even make them use appropriate distributions to do this and make them figure it out..... which is just evil!!!) 

    For example, I tell my students that a football team score and average of 2 touch downs and 2 field goals per game. Then generate 1,000 random samples for the number of touchdowns and field goals and tell me what is the average number of points scored by the team? (Assume a TD is 7 points and a FG is 3 points) How often is the team held scoreless? How often does the team score more than say 30 points? What is the maximum number of points they score? What is a bad assumption I made in this simulation and how would/could you fix it? etc.

    If you had them do something like this, you could have them plot the data too and look at how different binning schemes changes the visual aspects of the data. You could also have them compute the 95% CI for the data and compare those results to a 95% percentile interval (2.5%, 97.5%). And ask which one is best to use on that type of data?

    ------------------------------
    Andrew Ekstrom

    Statistician, Chemist, HPC Abuser;-)
    ------------------------------

    Original Message


  • 30.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-24-2020 09:36
    "In the "real world", "I don't know" is an acceptable answer.

    In at least the US, people don't like to admit they don't know something.  Maybe they fear appearing stupid, but pulling an answer out of the air only makes you appear stupid.  One experience: I asked what voting rule that some software was using, including a specific rule dating from the 18th Century.  The facilitator made some noises about it being "sophisticated."  After looking at the results, I confirmed that my intuition was correct.  I also saw how the developers threw in lots of useless statistics to create a false impression of their software.

    Having said "I don't know," what do you do?  If you really should know the answer, you're in a pickle.  Otherwise, if the question is relevant, offer to find an answer, whether through your own research or by contacting an expert.  Others can give better advice for irrelevant questions.

    A corollary is to not pretend to be an expert when you're not.  I had a coworker who was dispensing advice about a construct I developed without actually understanding it.  The lack of understanding was obvious because the construct was not appropriate for the context.  I recommended the in-house expert and followed up with a reference to a paper I happened to read later.


    ------------------------------
    Chuck Coleman
    ------------------------------

    Original Message


  • 31.  RE: What are your "real world" experiences at your jobs? (Guidance for students)

    Posted 03-24-2020 18:20
    Thanks for the very important reality check, Chuck! I think not having to be perfect lifts a weight off of students' back. Of course I am not trying to discount learning all one can, just that there can come a point where "I don't know" is an ok answer.

    I often forget, as I have all the book answers, that the "real world" sometimes doesn't have an answer, so collaboration and community are important components to any job.

    -Jennifer

    ------------------------------
    Jennifer Ward
    Clark College
    ------------------------------

    Original Message


Related Content