Sin-Ying Lin, PhD

PhD student @StonyBrookU.| Clinical psychology| Complexity of mental illness| Individual treatment trajectory| Advanced quantitative modeling

[Sin-Ying Lin, PhD @SinYing_Lin]

Does anyone know how to do time series on partially overlapping time frames? Each data point is a score averaged across the year, and a new score is published every quarter. No access to raw data. Appreciate any help!

[Sin-Ying Lin, PhD @SinYing_Lin]

Twitter friends, does any of you know anyone who works at Verily? I am interested in their data scientist position and feels like the job description matches my experiences well. I would love to get to know the job a bit more and potentially find an internal reference!

[Sin-Ying Lin, PhD @SinYing_Lin]

@Dr_Sayers @ryanwebler45 @dp_moriarity I feel the thousands of pre-internship practicum and externship hours are more than enough to conduct quality clinical research in certain areas. I have surely improved during the internship year but the huge financial sacrifice doesn’t justify the marginal amount of leaning gain

[Sin-Ying Lin, PhD @SinYing_Lin]

@JSchleiderPhD @IRIfellowship As Taiwanese, I wholeheartedly recommend Din Tai Fung! I’ve been to Maneki for Japanese food once. Also recommend it. For site viewing, Chihuly Garden and Glass is totally worth the ticket price!

[Sin-Ying Lin, PhD @SinYing_Lin]

@JSchleiderPhD @NorthwesternU @NUFeinbergMed Congrats! Jessica! This is amazing! They're so lucky to have you, and I'm sooo lucky to have the chance to work with you during my time at SBU!

[Sin-Ying Lin, PhD @SinYing_Lin]

@hawesmt Huge congrats 🍾🎈🎊🎉

[Sin-Ying Lin, PhD @SinYing_Lin]

@CraigAnthonyRS Just sent the link to all MIDAS RAs 🙌

[Sin-Ying Lin, PhD @SinYing_Lin]

@CraigAnthonyRS @DrBenKatz

[Sin-Ying Lin, PhD @SinYing_Lin]

Does any of you have a friend who is working as a behavioral scientist @ Apple (or any position along the line)? I'm very interested in this role and hope to build some connections. Appreciate any help!!!!!

[Sin-Ying Lin, PhD @SinYing_Lin]

@hwbchen Congrats! 🎉🎊❤️

[Sin-Ying Lin, PhD @SinYing_Lin]

@mcmullarkey @CraigAnthonyRS Thank you! Wonder if you could help review my resume and give some advice! Or any general advice would be helpful!

[Sin-Ying Lin, PhD @SinYing_Lin]

Academics: It's very easy to land an industry job with lots of transferable skills you obtain during your Ph.D.

Me with all the coding/statistical/quantitative/ML skills asked on the job post but without a CS degree: Got a rejection letter every other day 😃

[Sin-Ying Lin, PhD @SinYing_Lin]

Reading an oncology paper about integrating psych services with oncology treatment: "Unlike advanced medical trainees at a similar academic stage (medical residents), behavioral medicine externs are of little to no cost to the institution." Our free labor is very well recognized!

[Sin-Ying Lin, PhD @SinYing_Lin]

@hawesmt Yes, increased complexity definitely plays a role. There are teams rooting for complex models. Some even suggest always adding random slopes. Model complexity itself is not a sin. We just need to think about our sampling adequacy and pay close attention to model generalizability.

[Sin-Ying Lin, PhD @SinYing_Lin]

In theory, mixed-effects modeling for repeated measures should improve our model estimation accuracy. However, I tend to find that using a trained mixed-effects model for predictions on new samples performs worse than a regression model. (1/3)

[Sin-Ying Lin, PhD @SinYing_Lin]

Solution:
Only use a mixed-effects model for prediction purposes when you can use both the fixed- and random effects. For example, if you got students nested within schools and you were predicting incoming students' future performance in each of these schools. (3/3)

[Sin-Ying Lin, PhD @SinYing_Lin]

Potential Cause:
1. The random error terms are not truly independent (and in reality, they often correlate with each other), which violates the assumption
2. Model overfitting (2/3)

[Sin-Ying Lin, PhD @SinYing_Lin]

@hawesmt Yes. My CS, finance, economics, and stats friends can easily earn over $10,000 over the summer and it's their norm (but not a requirement) to do an internship. Well...it's our norm to do an externship for $0 and a mandatory internship for low wages.

[Sin-Ying Lin, PhD @SinYing_Lin]

Disclaimer: Not saying you can't do fabulous research with the software mentioned above. I personally know SPSP and SAS well too. Just saying it's going to be slower and more restrictive, especially when handling large-scale data.

[Sin-Ying Lin, PhD @SinYing_Lin]

Start looking for a data scientist job in Health Techs. Most of them claim to be innovative, and this is my reaction when I see them asking for experiences in SPSS, STATA, and SAS...Thought we were Gen Python/R???