The p-value is determined from the test statistic. The test statistic is calculated by using the sample mean/proportion, the sample variation (standard deviation) and the sample size. I should have mentioned above that my conclusions only hold if you assume only the standard deviation is the same fro the 600, 400 and of course the final 1000. Now in reality this is most likely not true, but given the relatively large sample sizes, you would assume the standard deviation to most likely be similar and any difference to be small and have a very little impact on p-values.
So to answer your question about whether you can predicting the p-value of 600 based on 1000, the simple answer is no because not only has the sample size changed but ultimately the standard deviation would change too. But if you know exactly what test was used, and assumed the standard deviations of the 600 and 1000 to be the same, then yes you could estimate the p-value of the 600 based on the 1000.
I'm not sure why you would want to predict the p-value of the 600 from the 1000 because in reality you could just crunch the numbers for the first 600 and come to the correct p-value. The reverse (trying to determine what would happen to the p-value of the 1000 based on the 600) would make more sense.
ASX Price at posting:
$41.35 Sentiment: None Disclosure: Not Held
