Friday, March 5, 2010

It's All About Timing (elevated from a comment to a post by Dr H)

Because this is an important issue, and I'm afraid Andi's thoughtful summarizing of the situation will be missed where it is, I'm copying/pasting/posting it here to get your attention.  The rest of this is all from Andi.  (also, don't miss Lynn's post right below this one, which I don't want you to miss because it is so COOL!)
I agree with Elizabeth - we should quickly come to a decision; until we do, our survey and data-collecting are at a stand-still. I also agree with Isaac in that we absolutely cannot use data from two separate years. The transportation and University-usage data must be from the same year or our results will be more skewed than they would be from estimating one or the other. We have three year choices: July 1, 2008 - June 31, 2009 June 1, 2009 - May 31, 2010 July 1, 2009 - June 31, 2010 Below, I've commented on what I see as pro's and con's for each. Please let me know what you think, and let's make a decision. July 1, 2008 - June 31, 2009 (+) True data for University-related electricity, waste, etc. (-) New student body - the transportation survey would have to extend only to the 2009-2010 sophomores, juniors, and seniors because the seniors from last year are no longer on campus, and the freshman of this year weren't at UWEC yet. This significantly narrows our participant pool. (-) Reliability statistically decreases the further back you go when surveys include self-report. In addition, the relevance for the student body would decrease. "Who cares about last year?" Though the information from last year IS important to us, it's the rest of the campus who has to feel like their responses will help make a difference. The more displaced they are from the dates/times in question, the less they'll be willing to "help." June 1, 2009 - May 31, 2010 (+) This keeps the same student body and keeps the survey relevant to participants. In addition (part negative, part positive,) the information will be entered into the CACPCC in early May (after spring break.) We would only need to estimate of the campus-collected information on electricity, waste, etc. for the month of May. My guess is that over the past few years, the waste, heating, etc. has been pretty stable for the month of May. Could we average and follow the trend? (-) Averaging May (-) This does not run with the school's fiscal year. How is the year separated? If it's just 1 month we're looking at, it's not a huge deal to add/edit that into a spreadsheet. If it's calculated annually, that would require taking each annual datum and dividing by 12. Still not hard to do, but it does increase error. July 1, 2009 - June 31, 2010 (+) Runs with the school year (+) Same student body (-) We would have to estimate two months worth of data (using past trends) which increases error. (-) Where would summer fit in for transportation? We could certainly separate out the summer months and ask for July/August, 2009 and then June, 2010, but that may be tedious for our participants. We could also average the summer months and add it into the spreadsheet that way, but again, it introduces further error. Other things to consider: Have any significant changes been made across our campus that would influence our numbers between 2008/2009 and 2009/2010? For example, has study abroad increased? Air travel is a large contributor to carbon or this campus. Big enough to matter? Has there been a greater push for recycling this year as compared to 2008-2009? Does that significantly influence our waste? Has the campus added electric/diesel fleets that would change the output between 2008-2010? When? How much do these things matter? I think we should take a vote. Pick one and explain why (as briefly as possible): Choice 1: July 1, 2008 - June 31, 2009 Choice 2: June 1, 2009 - May 31, 2010 Choice 3: July 1, 2009 - June 31, 2010

14 comments:

  1. My thoughts:
    I think estimating 1 or 2 months is workable.

    This is more relevent with current data; things change and we have been making changes our changes should be reflected since the last report.

    Estimating one or two months is fairly easy and maybe we shouldn't worry about being super accurate since your survey respondents will be estimating too.

    I believe there is value in 12 months worth of data even if some of the "data" is done by educated guesses.

    ReplyDelete
  2. My vote is for choice 2 (June 2009-May 2010). I agree with Lynn that estimating data for several months will not completely destroy the accuracy of our results, and I think it'd be best if we didn't split up the summer months. Furthermore, we're trying to measure the campus' current carbon footprint and I wouldn't want to use data that's not as current as we have available.

    ReplyDelete
  3. My primary vote would be for June 2008-June 2009. But if more people are set on using any of the 2009-2010 year, I would probably vote for the whole fiscal year. I'm pretty sure that the spreadsheet we are putting all this info into probably asks for things over the span of a year. Those who are unable to obtain information for the past few months can perhaps find out what happened in previous years and try to figure out an average.

    ReplyDelete
  4. Elizabeth, could you say a bit more about *why* you would vote for June 2008-June 2009?

    ReplyDelete
  5. I just think that a majority of the areas we need information from are ones that require raw data. Electricity, Heating, Waste, and Refrigerants all rely on what has happened in the year. It would be more difficult to try and estimate the data for the last few months of this fiscal year.
    But then I also realize that we run into a problem here with Transportation. I guess I feel that since the transportation section is based on a survey, we are already going to have information that is not totally accurate. I also feel that, from year to year, Transportation would be pretty similar (Most freshmen don't drive and live on campus, Sophomores might have more cars and some will still live on campus, and once you get to Juniors and Seniors, most will live off campus and have a car). So perhaps we can transpose our results from this year in transportation onto last year. This way we have the least amount of guess work. 4 sections will have very reliable and accurate information. Transportation will not be totally accurate, but as I said before, it's a survey, and I believe that transportation would be about the same.

    ReplyDelete
  6. My vote is for choice 3 (July 1, 2009-June 31,2010). I think it is important to use the most current data available to us that will most accurately reflect the changes that have been made since the last carbon inventory. I agree with Lynn that estimating two months worth of data is doable and not of great consequence. My guess is that data for May and June is fairly stable from year to year. I am not exactly sure how the summer months should be handled in the transportation survey, but I like the idea of the survey reflecting habits from this year. That way respondents will be more accurate in their answers and the results will reflect the current transportation trends, compared to the trends of last year, or the year before.

    ReplyDelete
  7. My vote is for choice 3. If estimating/forecasting the next couple months is not a big deal then it should be more accurate than missing an entire 1/4 of the student body on the survey.

    ReplyDelete
  8. I also vote for choice 3 (July 2009 - June 2010). I do not think that estimating some data will be incredibly challenging or will have a large negative effect on the reliability of our report. The positives of using information that is more recent and that includes changes that have been made to the university in the past year outweigh the negatives of estimating some data in my mind. I also think it is important to use a fiscal year. Though it may complicate the transportation survey a bit by splitting up the summer months it is important to make gathering data from administration as simple as possible. We want to make sure we keep a good working relationship with these various departments so they are more willing to help in future years.

    Using recent data for the survey is extremely important in my mind. (not just because I am on the transportation team!) As others stated before, if we try to collect information for the 2008-09 year we will be missing a large part of the population from that time - a population that I believe has a large impact on the carbon emissions on campus. (buses and personal vehicles) Also, if I were to take the survey and was asked to recall my habits from the 2008-09 year, I would honestly just think about this year and report that instead, because I would assume that they were somewhat similar and wouldn't want to take the time to think that far back. Essentially I feel that we will be getting 2009-10 data from the survey whether we ask for it or not...

    ReplyDelete
  9. I also agree with many that we should use choice three. No matter which option we go with, as has been pointed out, we will have negatives. Overall, I think the value of keeping our data current as well as relevant to our campus outweighs many of the downfalls that come with estimating.

    ReplyDelete
  10. I also agree that we should use choice three, the most recent fiscal year. I understand that it is not a perfect solution, but when you take into account all of the pros and cons, it will be the most useful/relevant for the purposes of the university. I do think that if we choose to do this one, we should come up with a standard way to estimate the months that we do not have data for (i.e. how many previous "Mays" will we average to get the data for the upcoming May?).

    ReplyDelete
  11. My vote is for June 2008-June 2009. I feel as if estimation is poor scientific method. Our campus has undergone a number of changes in recent years to lower emissions, and I think using past records to estimate levels today is illogical and discrediting. However, with transportation making up the majority of the footprint, I understand the importance of accurate results. I wouldn't be terribly upset if we ended up working with another time period.

    ReplyDelete
  12. I think we should go with choice 3 because I think it is important to include all current students. It seems that it would not be terribly difficult to estimate a couple months of data and would not be too big of a deal since our survey is not going to be 100% accurate anyway.

    ReplyDelete
  13. I agree with the consensus with choice 3 for the simple fact that everyone else iterated: it won't be hard to estimate a few months and we should have the most up-to-date information on changes were we can.

    ReplyDelete
  14. So I think we have basically made our decision but just to beat a dead horse...I vote for July 1st 2009 - June 30th 2010.

    I really wanted to choose choice 1 for its concrete data but I couldn't figure out a solution to the transportation dilemma.

    ReplyDelete

Note: Only a member of this blog may post a comment.