Could observed accelerating expansion be an illusion based on distance?

The thought occurred to me that if we just happened to be born into a time when the the universe started slowing down, it would take some time for us to observe this - especially in more distant galaxies. Locally, we might observe attractive or static systems of galaxies. But because the light takes so long to reach us, more distant galaxies would still appear to be speeding away or even accelerating while they could have actually slowed, stopped, or even reversed velocity! Is this possible? Are there other data to confirm the accelerating expansion of the universe that are fresher/closer?

• "actually" is a tricky word here. There is no global "now" for the whole universe, so what does this mean? – ACuriousMind Sep 17 '15 at 13:18
• @ACuriousMind: I think cosmological time is an acceptable proxy for a global "now". I am more worried about the headline asking if something that is an observed fact could be an illusion. I attribute that to the OP's unlucky choice of words since the body of the question is legitimate. I tried giving an answer but I find the situation too complicated to wing it without math. If cosmological time is a global phenomenon (i.e. the homogeneity assumption holds even in a strongly accelerating universe) there does not seem to be a simple answer to what the observed vs. actual acceleration is. – CuriousOne Sep 17 '15 at 14:01
• Apologies for choice of words here. When I say "illusion" I'm really just trying to convey that the events we see playing out when we observe distant galaxies happened long ago. – MW81 Sep 17 '15 at 20:55

If we want to observe the universe expansion with the most straightforward and direct way we have to measure redshifts of galaxies. The minimum distance where this effect would start to be observable is that at which the speed of recession is larger than the average noise speed, which is around a few hundred $km/s$. Given that the value of the hubble constant is around $70 (km/s)/Mpc$ galaxies should be at least of the order of $10Mpc$ away for speed of recession not to be overshadowed by the noise, which is around $30$ million light years, so that is how recent the observation would be, $30$ million years old (which is very up to date in cosmological terms). Notice that if the hubble constant was slower, then we'd had to look further away and further back in time.