Taureanbull yourself and others keep throwing around the term "statistically significant" and I have a strong feeling that most have no idea what that actually means. Not having a go at you, just hoping to make what I think is an important point for amateur investors to wrap their head around. I am NOT going to claim to be any sort of expert in this area. I did well in statistics at uni with very little study (health degree so I didn't place much importance on stats) however have barely used it since. Apologies if I make some mistakes:
So basically statistical significance refers to whether the results seen in your sample are likely to yield similar results across an entire population.
For gastric cancer, given ROUGHLY 26K new cases are diagnosed each year, this will be my population. Our data set is a sample of 10 patients. The significance level is usually set at 5% or lower, so to err on the conservative side (ie reducing the size of the sample size required) I will use 5%. I will also use a confidence level of 95%.
With the figures above, the minimum sample size required in order to achieve a statistically significant result is 379 patients. On the other side of the coin, with a sample of 10 patients, our margin of error is a massive 31%. In other words, to try and extrapolate from our study: how many patients might be expected to experience a complete response for example, this would be 10% +/- 31%, or in terms of our gastric cancer population 2600 +/- 8060. Furthermore, decreasing the margin of error has a large effect on the sample size required. For example dropping the margin of error to 4% rather than 5% results in minimum sample size of 587. A 99% confidence level with a 2% margin of error would need a sample size of 3587.
As you can see, the sample size we currently have is simply too small. That's not to say the results aren't exciting, clinically significant, promising etc. Just not statistically significant IMHO.
IMU Price at posting:
1.9¢ Sentiment: Buy Disclosure: Held
(20min delay)
