In cancer research, is it better to study if the modulation of a lncRNA is able to change the tumor size at different times or to see a change in mice survival?
tumor size monitoring seems to be more straightforward, in my opinion. Mice survival can be affected by many other factors (bias) and high population sample (number of animals) should be necessary.
Even better way could be measurement of some biochemical markers at defined time points or doing some molecular analyses (e. g. whole transcriptome sequencing to see the molecular pathways which were affected by your treatment).
I would suggest to check both parameters. Of course, tumor size is better, as by measuring it you will know the direct effect on tumor growth, but survival could be also useful if you have a very aggressive tumor model. Martin is right, you also should include molecular analysis of your samples.