@scienceandbears In one sense there’s nothing to negotiate (unless postdocs are exempt from the new rules) 38k is the new minimum salary for international post docs. Think pis need to think more carefully about the pay and conditions they offer for a job that requires at least 8 years of training
@robinsall I don’t think it was suggesting it’s a failure but highlighting socioeconomic background continues to be a significant predictor of success even with a degree. But what’s happens vast majority? Because the rich kids probably aren’t the ones doing creative arts at the ex-polytech
@robinsall While in an ideal world people could attend university for the sake of the education alone , it is instead a significant financial decision. Even more so if you are form a lower socioeconomic background
@robinsall I attended a school whose intake areas are amongst the most deprived in the country. The message was uni leaves you on average better off. For many this is not they case and they end up in jobs that didn’t require a degree but now with an effective ~9% higher tax rate
@AnhHLe2702 So if you were to do the “do the best-fit of selected unshared parameters differ” it would do what you’re looking for
https://t.co/cp8UqxYRwD
@AnhHLe2702 Maybe I’m misunderstanding, but practically doesn’t something like the compare tab for curve fitting in prism do exactly what you want to do ?
https://t.co/Rox64KfvV8
@AnhHLe2702 Not sure I follow? Would you not want to know instead something like statistically does the best fit value for a parameter differ between the datasets, Or are both datasets explained by the same curve? You know the 2 curves are different, it’s about how they relate to your data
@AnhHLe2702 If the fitted curves can’t be translated / make sense with regard to your hypothesis , then the question would be why are you fitting that curve to your data ?
@AnhHLe2702 A connecting line between the two values is only true for the sem , an sd and 95% ci would extend past the two values. I would argue an error bar is never really a good indicator of variability , if this what you want to show you should show all points / do a box and whisker etc
@AnhHLe2702 @RossTAPedersen Having standards of diluted sample is probably good enough ( although most people don’t do this because they want to load more exp samples). You just want to be sure that differences you’re seeing are not just pipetting error exaggerated by poor dynamic range
@AnhHLe2702 @RossTAPedersen Plus who knows what the dynamic range of protein detection actually is ( although in about 10 years I’ve only seen someone try to figure this out once 🙈)