core Q11 - seasonal or not?

[miss_muffiny]

after coming back to this question and staring at it for a good 10 or so minutes, i changed my answer from E to D. not sure if that was a good move or not.  i think this one is aaaaaaaaaalmost 50/50 either way. its open to interpretation. there is definately a cycle of every 6 months, but is that enough/too much to be seasonal??

not meaning to be nit-picky, just interested in everyone elses opinions

[RD]

I left mine as E.

[leok]

I reckon it was D

[ilovesuck]

i think the general consensus is that it is E.

(i put D )

[Noblesse]

I chose D, but honestly have no idea. I thought it was seasonal enough but I guess not :/

[leok]

Na if you look carefully the months that the trend increases is random

[*Roxxii*]

I chose D, after about 10 minutes of stressing and switching back and forth in the exam.
But as soon as i got out of the exam room and could think clearly, i realised it is most likely E.

For a time series to be seasonal, there needs to be a regular time interval between each peak.
Yes, you can ASSUME that there is a regular cycle of every 6 months, peaking every May and Nov (month 5 and 11)
But we are only given one year of data, and there is only ONE interval between the peaks.
There is no indication how many months there will be between the peak at month 11 and the next peak (if there is a peak at all)
Thus you cannot come to the conclusion that it is seasonal with the information given to us.

If there were AT LEAST 3 peaks, so 2 intervals in between...and the intervals had the same number of months, then it would be reasonable to assume the trend is seasonal.

[costargh]

I rekon its E because just looking at the type of thing it is. "Online help service"

Why would there be any seasonal trend with that?

Fair enough if it was for a fruit picker or christmas hampers, but for online help? There would be no logic behind why an online help service would have a seasonal trend. Perhaps if you wanted to make a loose connection you could say that every July-Aug (i think) when people get some of their tax back from the ATO which causes them to have more discretionary income which means a higher possibility of people buying a computer of laptop in the coming month which would lead to higher internet subscriptions and therefore more online help services... but thats bery unlikely. Plus that would require more than one year to compare
Btw I put D. Stupid me.

[shonakennedy]

you guys make me laugh. it is not seasonal if there are only trendsshown for one year!!!!
if it were seasonal, the graph would have depicted results from multiple seasons, not just one single year.

[costargh]

Condescending much. Geez

[shonakennedy]

ah crap i didnt mean to be im sorry.

[jsimmo]

If q11 had a seasonal variation it would be highly likely that it would be followed by a seasonal indice question... yet question 13 was do with least squares regression.. which leads to consider that there was no seasonality involved.

[*Roxxii*]

If q11 had a seasonal variation it would be highly likely that it would be followed by a seasonal indice question... yet question 13 was do with least squares regression.. which leads to consider that there was no seasonality involved.

Nicee
Very good observation there jayden 

I like the way you think!!

[SP3CIALIS7]

If q11 had a seasonal variation it would be highly likely that it would be followed by a seasonal indice question... yet question 13 was do with least squares regression.. which leads to consider that there was no seasonality involved.

Nicee
Very good observation there jayden 

I like the way you think!!

Yeah, the fact that there was no de/seasonilsation in the multi-choice makes me think it's very likely it will feature significantly in short answer

[miss_muffiny]

ok so based on all that pretty sure the answer was E, and yet most of us put D. danggit!!


Yeah, the fact that there was no de/seasonilsation in the multi-choice makes me think it's very likely it will feature significantly in short answer
 

...good point. study up deseasonalising!! oh and there were no transformations either. or contour maps, or break even analysis, or stem and leaf plots... hmm!!