Week 9: Finishing Touches
Hey everyone, welcome back to my blog! This week I decided to look at potential errors or biases in the data and models and what can be done to eliminate or reduce these. Additionally, based on findings from the models and research and survey, I wanted to create a list of what companies can do to support the mental health of their employees.
Errors, Biases, And Limitations:
One limitation of the models is how it does not take race/ethnicity into account, since there was no race/ethnicity question in the dataset. If this was one of the questions, it could help determine if an employee’s race/ethnicity plays a role in their mental health or makes them more or less likely to receive treatment. Since there is no race/ethnicity column, we do not know if the data over-represents one race/ethnicity, so it might be inaccurate when applied to real tech companies or a real population of tech employees.
Another limitation of the models is that since they use an optional online survey, the data might be biased. The employees who took this online survey may not represent the entire population of tech employees, since a tech employee was never defined, so if this project was applied to the entire population, it may not be accurate. Thus if I were to continue the project, a next step could be to collect surveys from a variety of demographics and tech companies, to see if these findings would be different for different populations. Additionally, more research can be done to determine what might be some causes of different demographics being more or less likely to require treatment.
The models could be used as an estimation for determining how many and which employees may require mental health interventions, but in the end, companies should provide ample resources for all of their employees. We saw other limitations in these models: the logistic regression model and neural network gave a significant number of false negatives which are dangerous because a company may under-allocate resources based on their predictions. Both of these ethical considerations show that the model should only be used to get a baseline estimate of the amount of resources needed.
What Can Companies Do:
The tricky part of trying to answer this question is that each company is different, and they all may be providing different levels of support.
Before I created a list, I decided to look at the suggestions employees gave in my survey, so that I could include those:
- “Wellness Day Off”
- “Periodic Health Check”
- “Encourage To Take Frequent Breaks”
- “Conducting Monthly Mental Health Exposure Sessions, Pointing To The Right Resources And Providing Onsite Counselors To Whom We Can Open Up And Discuss Mental Health Issues.”
- “Inclusion Of Mental Health Discussion Can Be Incorporated Into The Performance Appraisal As Well As Regular Discussions Between The Supervisor And Employee.”
- “By Enforcing Auto Logout From System .”
- “Provide Better Work Life Balance. Reduce Work Hours.”
- “It Would Be Great To Have An On-Site Counselor(S) In Medium-Bigger Organization. It Would Be Like Visiting Psychiatrist”
- “Provide Some Physical Activities Like Yoga Classes Or Should Have Gym To Give Chance To Employees For Doing Something Else Then Regular Work”
- “Arrange Yoga And Meditation Classes.”
- “Employers Should Make Sure Leaders Within The Company Understand The Importance Of Mental Health. Make Them Undergo Training Needed From Time To Time”
- “Ensure That Staff Feel Comfortable To Take Mental Health Type Vacation Days Regardless Of Workload At Relatively Short Notice. There Is Always A Lot Of Work To Be Done And Staff Feel Like They Can’t Take Time Off To Rest And Recharge Because Of It.”
This list that I have created is through previous research and general observations from the survey that I sent out. This is the bare minimum that companies should definitely do, not a comprehensive list of everything they can do.
Companies should do the following:
- Address The Stigma Around Mental Health Issues And Raise Awareness About Them
- Normalize Policies For Requesting Time Off For Mental Health, Despite The Amount Of Workload Present, Just Like Sick Leave Or Vacation Days
- Spread Awareness About Their Provided Mental Health Programs And Emphasize Their Anonymity, So That Employees Feel Comfortable Taking Advantage Of These Resources
- Hire A Company Mental Health Counselor (Maybe More Than One If Needed) And Encourage Employees To Talk About Their Mental Health With The Counselors
Again, this list is definitely not comprehensive, but it is the minimum that companies should do to invest in improving the mental health of their employees.
Next week, I will talk about another approach tech companies can take, and I will begin creating one last survey for employees to take, which will measure how effective these solutions could be.
Thank you for reading!
- Open Sourcing Mental Illness, LTD. “OSMI Mental Health In Tech Survey 2016.” Kaggle.Com, 2016, Www.Kaggle.Com/Datasets/Osmi/Mental-Health-In-Tech-2016.