• AT Hiking Rates, Section by Section (AT Data and Schedules)

    AT Hiking Rates, Section by Section
    (updated in February 2011)

    by map man (Steve Shuman)

    I often see anecdotal estimates of how long it takes to hike various sections of the AT, sometimes as miles per day estimates and sometimes as proportions (someone might say, "allow three days to hike in the White Mountains a distance that had taken two on other parts of the trail"), but since the advent of Trailjournals.com we now have access to hundreds of detailed accounts of AT thru-hikes, and since I'm a numbers nerd from way back I decided to see if I could verify with real numbers just how long the typical Thorough Journal Keeper (I'll call them TJKs from now on) takes to hike the AT and various sections within the AT.

    So I went through many, many journals day by day for the 2001 thru 2010 hiking seasons and ended up with a group of 240 TJKs who did such a precise job of documenting not only when they started and stopped their thru-hike and when they passed certain relevant landmarks along the way, but even accounted for when and where they took their "zero days."

    This study is limited to north bound thru-hikers (NOBOs) who completed their hike in one hiking season and passed ten section defining landmarks along the way in south to north chronological order (though the study does include hikers taking occasional SOBO dayhikes within sections). The sample sizes would not be large enough yet for me to make a meaningful study of SOBOs or flip-floppers, but someday I would like to replicate this study for SOBOs.

    The section defining landmarks I chose either have psychological significance (Harpers Ferry is an example of this -- it's often called the psychological half-way point) or mark a change in topography that can influence hiking rates (the Glencliff to Gorham section is an example of this). The points I chose are: Georgia Border, Fontana NC, Damascus VA, Waynesboro VA, Harpers Ferry WV, Delaware Water Gap (DWG) PA, Kent CT, Glencliff NH, Gorham NH and Stratton ME. Combined with Springer and Katahdin they mark off eleven distinct sections.

    The study includes 7 NOBOs from the class of 2001, 17 from 2002, 24 from 2003, 33 from 2004, 24 from 2005, 38 from 2006, 30 from 2007, 25 from 2008, 25 from 2009 and 17 from 2010 (these were the only journals from these years detailed enough for this study). This is how the journal keepers broke down by gender: 165 were male, 42 were female, 31 journals were for a male and female hiking together, 1 was for two females hiking together and 1 was for two males hiking together (I count each journal as one hike for this study even if it is for multiple people). When calculating the distances between my landmark points I was aware that these distances sometimes changed slightly from year to year as the trail changes, so with the help of the AT Data Books for various years I calculated a weighted average for the distance of each section based on how many hikers were in the study for a given year.

    The result of all this journal reading and number calculating is a DESCRIPTIVE study of a certain thru-hiking population (NOBO thorough journal keepers) and may or may not be representative of all thru-hikers (though I hope it's pretty close). This study is in no way meant to be a PRESCRIPTIVE analysis of how people OUGHT to hike the AT.

    So with that said, let's get to the good stuff. Table 1 shows the average (mean) number of days it took these TJKs to hike each section. The first number is the days for that section, the second number is the running total for the hike, and the median number of days to hike each section is listed after the description of that section (The mean number of days for these TJKs to thru-hike was 168.8 while the median number was 171.):


    TABLE 1 -- Days to Complete Various Sections

    DAYS ~~~ TOTAL DAYS ~~ SECTION
    8.0 days..........(8.0)............Springer to Georgia Border (7.7 days)
    7.9 days.........(15.9)...........Georgia Border to Fontana (7.7 days)
    24.4 days.......(40.3)...........Fontana to Damascus (24 days)
    28.7 days.......(69.0)...........Damascus to Waynesboro (28 days)
    11.2 days.......(80.2)...........Waynesboro to Harpers Ferry (11 days)
    19.2 days.......(99.3)...........Harpers Ferry to DWG (19 days)
    12.6 days......(111.9)...........DWG to Kent (12 days)
    23.5 days......(135.4)...........Kent to Glencliff (23.2 days)
    9.7 days........(145.1)..........Glencliff to Gorham (10 days)
    9.9 days........(155.0)..........Gorham to Stratton (9.85 days)
    13.7 days...... (168.8)..........Stratton to Katahdin (13.6 days)


    When I started this study I was not going to try to figure out how many zero days were being taken but it quickly became apparent that some sections were a lot more prone to hikers taking zero days than others. This would in turn have an effect on how long it took to hike each section and might provide somewhat misleading numbers when calculating miles per day for the average TJK in any given section. Did a section take a longer time to hike solely because of difficulty or were many tempting places to take zero days in the section also playing a role? By figuring out how many zero days were taken in each section both Miles Per Day (MPD) and Miles Per Hiking Day (MPHD) could be calculated. When stripping out the zero days from the calculations, Table 2 shows a remarkably smooth linear progression in the number of miles covered in the first four sections of the trail as TJKs gradually increased the number of miles hiked per day (in the Miles Per Hiking Day calculation). The table also shows that thru-hikers were slowed down by the rugged terrain of the White Mountains and western Maine, though perhaps not quite as much as legend suggests. The first number in this table is MPD (Miles Per Day) and the second number is MPHD (Miles Per Hiking Day). The weighted distances for each section follow the section description:


    TABLE 2 -- Miles Per Day and Miles Per Hiking Day

    MPD ~~~~~~~ MPHD ~~~~~~ SECTION
    9.4 miles..........(10.1 miles).........Springer to Georgia Border (75.6 miles)
    11.2 miles........(12.0 miles).........Georgia Border to Fontana (87.5 miles)
    12.2 miles........(14.0 miles).........Fontana to Damascus (297.1 miles)
    13.4 miles........(15.9 miles).........Damascus to Waynesboro (388.6 miles)
    14.4 miles........(16.8 miles).........Waynesboro to Harpers Ferry (161.1 miles)
    13.9 miles........(16.8 miles).........Harpers Ferry to DWG (270.3 miles)
    13.9 miles........(16.1 miles).........DWG to Kent (172.4 miles)
    14.0 miles........(15.5 miles).........Kent to Glencliff (323.8 miles)
    10.5 miles........(11.4 miles).........Glencliff to Gorham (100.6 miles)
    11.1 miles........(12.5 miles).........Gorham to Stratton (110.1 miles)
    13.6 miles........(14.7 miles).........Stratton to Katahdin (187.9 miles)
    12.9 miles........(14.7 miles).........The entire AT (2175.0 miles)


    Here's the distribution of hikers grouped by the month they left Springer and the month they reached Katahdin:

    2 hikers in this study left Springer in January
    30 in February
    144 in March (60%)
    57 in April
    7 in May

    4 arrived at Katahdin in June
    35 in July
    52 in August
    113 in September (47%)
    36 in October

    The first date in Table 3 is the median date each point was reached by the TJKs (same number of hikers arriving after this moment as before) and the second date is the mean date:


    TABLE 3 -- Date Landmarks Were Reached

    MEDIAN DAY ~~ MEAN DAY ~~~ LANDMARK
    March 17............March 20...........Springer
    March 25............March 28...........Georgia Border
    April 1................April 5...............Fontana
    April 28..............April 29..............Damascus
    May 29..............May 28..............Waynesboro
    June 9...............June 8...............Harpers Ferry
    June 29.............June 27..............DWG
    July 12..............July 10...............Kent
    Aug. 5...............Aug. 2...............Glencliff
    Aug. 15.............Aug. 12..............Gorham
    Aug. 26.............Aug. 22..............Stratton
    Sept. 9..............Sept. 4..............Katahdin!


    Of course no single hiker is "typical" and people will vary in their own ways from the 168.8 days (about five and a half months) it took these TJKs to get to Katahdin. But I think it can be useful to see the rate of progress these TJK thru-hikers have experienced on their way there for planning purposes. One thing I discovered is that if a group of hikers in this study taking four months to thru-hike takes 20% less time to get there than a group taking five months, 33% less time than six month hikers and 43% less time than seven month hikers, those figures tend to stay true for each section within the thru-hike as well. So with that in mind I calculated the number of days a "typical" TJK thru-hiker might have needed to reach these section landmarks for four different hypothetical hikes. Table 4 lists the "typical" number of days it would take to reach each landmark for hikers taking 4 months (122 days), 5 months (153 days), 6 months (183 days) and 7 months (214 days) to thru-hike:


    TABLE 4 -- Four Hypothetical Hikes

    4#HIKE ~~~ 5#HIKE ~~~ 6#HIKE ~~~ 7#HIKE ~~~ LANDMARK
    6 days...........7 days..........9 days..........10 days..........Georgia Border
    11 days.........14 days........17 days.........20 days..........Fontana
    29 days.........37 days........44 days.........51 days..........Damascus
    50 days.........63 days........75 days.........87 days..........Waynesboro
    58 days.........73 days........87 days.........102 days.........Harpers Ferry
    72 days.........90 days........108 days.......126 days.........DWG
    81 days.........101 days......121 days.......142 days.........Kent
    98 days.........123 days......147 days.......172 days.........Glencliff
    105 days.......132 days......157 days.......184 days.........Gorham
    112 days.......141 days......168 days.......197 days.........Stratton
    122 days.......153 days......183 days.......214 days.........Katahdin


    Finally, I wanted to take a close look at the nature of "zero days," the days that no miles are logged by hikers on the AT (TJKs took a mean 20.7 of them on their thru-hikes -- the median number of zero days was 19). I wanted to look at both the short term breaks (1 or 2 day breaks from the trail) and long term breaks (3 or more consecutive days with no AT miles hiked). I hypothesized that hikers would need to take a lot of the short term breaks in their earlier days on the trail to cope with hiker's fatigue and with the sometimes nasty weather in the southern Appalachians in March and April, and that these short term breaks would lessen in frequency as a hiker walked north. It appears I was wrong, as Table 5 shows. TJKs took very few of these zero days in the first two sections. It's my speculation now that hikers seemed to take these short term breaks largely due to the availability of trail towns and concentration of hiker focused shuttle services and hostels. For example, the section with the highest percentage of these short term zero days taken was the Fontana to Damascus section with Hot Springs, Erwin and many famously hospitable hiker services in this stretch.

    On the other hand, I hypothesized that long term breaks (breaks of 3 or more consecutive days when hikers often leave the vicinity of the trail completely) would be scarce in the early days when the novelty and newness of the experience alone might carry people forward and again scarce toward the end when the goal was so close, and more frequent in the middle of the journey. On this, it sure looks like I was right, as the "Long Term Break" percentages in each hiking section in Table 5 show. In this table the first number is percentage of days taken to complete a section that are zero days. The second and third numbers break the zero days into two groups -- STBs (zero days taken in Short Term Breaks of 1 or 2 days) and LTBs (days taken in Long Term Breaks of 3 straight days or more):


    TABLE 5 -- Zero Days

    %ZERO DAYS ~ %STB ~~~ %LTB ~~~ SECTION
    ....(5.7%)..........(4.8%)........(0.9%)........Sp ringer to Georgia Border
    ....(7.5%)..........(6.3%)........(1.2%)........Ge orgia Border to Fontana
    ....(13.0%)........(9.0%)........(4.0%).........Fo ntana to Damascus
    ....(15.0%)........(8.2%)........(6.7%).........Da mascus to Waynesboro
    ....(14.3%)........(7.8%)........(6.4%).........Wa ynesboro to Harpers Ferry
    ....(16.3%)........(7.3%)........(9.0%).........Ha rpers Ferry to DWG
    ....(14.9%)........(7.6%)........(7.3%).........DW G to Kent
    ....(11.1%)........(7.2%)........(3.8%).........Ke nt to Glencliff
    ....(9.1%)..........(6.8%)........(2.3%).........G lencliff to Gorham
    ....(11.1%)........(8.5%)........(2.6%).........Go rham to Stratton
    ....(7.3%)..........(5.8%)........(1.6%)........St ratton to Katahdin
    ....(12.3%)........(7.5%)........(4.8%).........Fo r entire AT


    METHODOLOGY

    If a hiker started on the approach trail to Springer and only went as far as the Springer Mountain Shelter I didn't count that as the first day of the thru-hike even though .2 miles of the AT were covered. Likewise, if a person got a ride to USFS 42 and walked the .9 miles to Springer and hiked no more of the AT that day I didn't count that as the first day of the hike either. I think these small partial days would distort the results for the first short section to the Georgia border. The day a hiker passes USFS 42 going north is the day I start the thru-hike clock ticking for the purposes of this study.

    When a hiker reached one of my landmark points -- for example, Waynesboro -- I stop the clock for that section (in this case, the Damascus to Waynesboro section) and start the clock for the next section. So any zero days that hiker took in Waynesboro are counted in the Waynesboro to Harpers Ferry section.

    If a hiker passed a landmark, let's say DWG, and hiked on past without stopping for the day, I break that day into fractions of tenths of a day. So if that day began at Kirkbridge Shelter, 6.4 miles short of DWG, and ended at the "Backpacker Site" 4.8 miles past DWG, I counted six tenths of that day in the Harpers Ferry to DWG section and four tenths of that day in the DWG to Kent section.

    In tracing hikers' progress in their journals I used all the clues available to tally zero days and hiking progress. Some were very thorough and gave exact starting and ending points for each day, with mileage accurately logged and separate entries for each zero day as well. These journals were easy to follow. But not all journals used in the study were this thorough. Some just gave starting and stopping points. Some only registered mileage. Some were odd combinations of the two. Some left gaps when they took zero days. Some would recount multiple days of hikes in one entry (and all of these oddities often meant that the "Stats" section available to look at for each journal at Trailjournals.com had inaccurate numbers for "zero days" and "hiking days"). As long as I could reconstruct what had happened, even if it took reading the entire text of multiple journal entries to get it done, I made every effort to do it. But if there was anything in the journal that made me uncertain if every stretch of trail was actually hiked, and about tracking which days were hiked, and which devoted to zero days, I did not include that journal in this study.

    (For a more in-depth discussion of how data was gathered for this article, and a series of tables and illustrations going into more detail about different aspects of the data, as well as my responses to suggestions from White Blaze members with more knowledge of statistical methods than I have, see Post #28 in this thread. For information on which towns TJKs were most likely to take zero days, see Post #69. For preliminary findings on how the numbers for men and women compare, see Post #80. For a table comparing miles hiked per day with trail ruggedness, see Post #93.)


    ACKNOWLEDGEMENTS

    A big thank you to the folks who created and maintain Trailjournals.com and WhiteBlaze.net. These sites are a great service to the hiking community. I could not have done this study without Trailjournals and could not hope to share it with so many people without WhiteBlaze. And thank you as well to all the WhiteBlaze members who offered suggestions to improve this article or offered encouragement.
    This article was originally published in forum thread: AT Hiking Rates, Section by Section (AT Data and Schedules) started by map man View original post
    Comments 115 Comments
    1. Roland's Avatar
      Roland -
      Wow! How incredibly time consuming this must have been. I'm going to have to read through it a few times to fully appreciate all the information you present.
    1. Topcat's Avatar
      Topcat -
      Map Man,
      I hate to mooch others work but i would love to see the raw data on this study. I do statistics for work and also teach it. This would be great for me to use, something different and interesting compared to the dry stuff in our curriculum. Thanks for the interesting analysis.
    1. Kerosene's Avatar
      Kerosene -
      Whoa, someone's got a little bit of extra time on his hands!

      This is quite interesting, map_man. I'll bet that the ATC might be interested in your analysis.

      When all is said and done, though, I hope that no one (except potential record-setters) use this to modify their thru-hike. A lot of the pre-planning goes out the window when you get an injury, hit a big storm, get off the trail with your new-found buds, forget that the post office is closed on Sunday, etc. They will, however, be able to classify themselves in retrospect and see how they stack up to "Joe Hiker".
    1. Cuffs's Avatar
      Cuffs -
      That is one piece of phenomenal work! I do have a couple questions for you...

      Do you have the numbers on the genders? And what about the age brackets?

      I read many many trailjournals, and those are the 2 things I look for. Looking to find people (women) in my age group (35-40) that are doing a thru. I find I can understand and empathize with them and learn alot from them.
    1. Whistler's Avatar
      Whistler -
      Okay. Wow. What a cool study. Let's put this one in the Articles section. Great work, map man.

      -Mark
    1. Billygoatbritt's Avatar
      Billygoatbritt -
      Simply awesome!
    1. attroll's Avatar
      attroll -
      Wow, Oh my godness. I can not imagine how much time was spent on this. Maybe this thread should be moved to the Articles section. What do you think Doctari?
    1. Tha Wookie's Avatar
      Tha Wookie -
      I think your right, Atroll. This is a good piece of work. Still, I'd like to hear more about the methods.

      How were the averages weighted?

      What did the distributions look like? Normal? How did you deal with outliers?

      VERY COOL analysis, map man.... thanks for sharing your work.
    1. Alligator's Avatar
      Alligator -
      Some thoughts.

      The TJK's were grouped without regard to start date, start year. It hasn't been demonstrated that this is a reasonable assumption. There could be year-to-year mean differences, or even start month-to-month differences that have been masked by taking an overall grand mean. Further, this grand mean may be biased if there are significant year-to-year or start month-to-month differences.

      Kudos for defining the sample population. It is a self-selected sample, not a random sample however. That bothers me, not necessarily anyone else. I will say though that while everyone who starts hopes to finish, the reality is that about 80% don't.
      The result of all this journal reading and number calculating is a DESCRIPTIVE study of a certain thru-hiking population (NOBO thorough journal keepers) and may or may not be representative of all thru-hikers (though I suspect it's probably close).
      It is also stated that journalers with ambiguous zero days were dropped, another subsetting. Just splitting hairs here .


      But I think it can be useful for some hikers who are so inclined to get an idea of the rate of progress a typical hiker might experience on their way there.
      I may be taking this out of context but ... the typical hiker fails. About 80-85% of them. The population that these numbers may apply to are successful thru-hikers. Drawing inferences is not possible, because starting out, no hiker really knows if they are going to finish. This is a descriptive study, as stated.


      Regarding zero days. On average, by the results presented, it takes about 16 days to get to Fontana. IMO, that may not be enough time before the long reality of the journey sets in. In other words, injuries may not have seriously developed yet, mental fatigue may have yet to develop, the weather could still go bad, etc. I don't feel like the data is sufficient for the following.
      After having read a fair number of journals for my own enjoyment I hypothesized that hikers would need to take a lot of the short term breaks in their earlier days on the trail to cope with hiker's fatigue and with the sometimes nasty weather in the southern Appalachians in March and April, and that these short term breaks would lessen in frequency as a hiker walked north. Turns out I was dead wrong, as Table 5 shows.
      The data could be available. The journals might give a reasonable idea as to the purpose of the zero day. "I was nursing a sore tendon", "I needed a beer", "I had to make repairs", etc.

      Of course, standard errors or confidence intervals would greatly improve the understanding of the means . There is no description of the variability of these estimates.

      A very detailed analysis though and I applaud the effort .
    1. ARambler's Avatar
      ARambler -
      Quote Originally Posted by Alligator View Post
      Some thoughts.
      ...
      A very detailed analysis though and I applaud the effort .
      Actually, I didn't understand may of the thoughts in between.
      For me it is very important to look at only data for those who completed the trail.

      I recalculated Table 4 for my 2004 hike. I took the number of days in Table 1, multiplied them by the days hiked (100%-%zero)/100 and adjusted by the (total non-zero/my nonzero days). Even though the number of days I hike was relatively far from the mean, I was only off from these calculated days by +/- 1 day. Also, almost all of the discrepancy can be explained by snow around Franklin and slowing down for the last section between Stratton and Katahdin. I was over a day faster for these two sections in 2005. (On the other hand I lost a couple days for snow near Erwin and a couple days for tendonitus in NJ in 2005)

      Was there a greater variation in non zero day pace for this last section? It seemed to me that half the hikers were putting their head down and charging to Katahdin and the other half were trying to keep the hike from ending.
      Rambler
    1. dje97001's Avatar
      dje97001 -
      Alligator, the value is in that looking at hikers who finish the whole thing may provide lessons as far as pacing (esp. at the beginning when you are worried about people passing you, or wondering whether you started a week or two too late to make it to ME in time) and setting more realistic approximations for mail-drop locations.

      It might be interesting to examine those who don't finish and compare daily mileage... maybe they pushed themselves too hard too fast.... maybe they realized that they would never make it to maine at the speed they were capable of... all of it is interesting stuff.

      Yes, it is obvious this isn't a predictive model... but who cares? Consider the years he examined to be the population that exists. The means for the population may be significantly different than the means for those who don't complete the thru, those who don't journal online at trailjournals.com, those who completed in a previous year (not included in the analysis), or those who completed the hike by sections, etc. ... but he didn't claim this extended to those pops. Still, to worry about kurtosis or skewness in this case is pretty much a waste of time--most people don't even bother checking for that stuff anyway... they just live and die by the central limit theorem. Sure, an exceptionally rainy year may have slowed people early on (resulting in more zero days) but then really dry years may have resulted in faster paces (with fewer zeros). It should all even out (life is all about probabilities). We're only talking about 5 years (and ideally we'd have more) so there is likely to be a larger SE than you'd like, but without anything else to use, this is damn good stuff. You could always compare the 3 measures of central tendency to get a better idea of whether or not you have some outliers screwing with the data if you are really worried about it, but again, why bother, this is really interesting to chew on. Thanks map man!
    1. Peaks's Avatar
      Peaks -
      Lot of great work. Thanks Map Man.

      Roland Mueser did a limited survey in 1988 of thru-hikers. His survey shows a mean of 174 days, and 24 zero days average. I'd say it's a close correlation. So, while many things have changed, your analysis shows that some things have not changed.
    1. Tha Wookie's Avatar
      Tha Wookie -
      I think the last two folks who've responded to Alligator have missed his point that the data is good, but there really is no way of knowing how close to reality it is without understanding the nature of the data set. If you have a heavily skewed distribution, then mean averages can be very misleading. In those cases, the outliers could be dropped, the data could be weighted (he already said they were, but by what factor?) or the medians used in place of means.

      Like what gater said, the sample population is what it is. It's like all the psychology studies that can only be extrapolated to college students, becuase they were the sample group.

      All in all, valiant effort, and I think it can be made better to really solidify the results. Alligator isn't just being picky, but adhering the assumptions of stats (normal distribution, random sample, ect.).

      Interesting conversation.
    1. camich's Avatar
      camich -
      I think this is great. Thanks for all the time you put into it. I'm always happy to have additional information to help plan.
    1. dje97001's Avatar
      dje97001 -
      I get all of that. I think we all do. But you know as well as I do that very few studies actually use anything other than convenience samples (which are non-random--I'm not talking about random assignment here) because of cost, time and difficulty in compiling the true pop list. So basically we all assume normality, again unless we look at the skewness (or how flat or peaked the distribution is), which no one does. I'm sure you could do it, but again, I'm not sure what good it would do. Just compare the mean to the median... the closer they are to eachother the less likely skew exists.

      But let's be honest, we aren't doing significance tests on this data, nor ANOVAs nor Correlations nor anything else for that matter. If you wanted a study that could be published in a journal you probably want to worry about these things--yet again, this is a content analysis not necessarily subject to the same issues of experimental research (samples for CA are often non-random). Frankly with a sample size of 105 (unless map man included Squeaky) there aren't likely to be substantial outliers (again, these people are already outliers from the "normal" population... most people wouldn't walk over 2000 miles). I think that this compilation of data is awesome in and of itself and really doesn't need anything else.

      I probably over-reacted, but Academics (who have had sufficient training in stats and methodology) commonly do what Alligator did: finding potential flaws/holes and making it apparent that if the flaws did exist then any conclusions would be shaky at best and then finishing it up by saying something nice about the effort--no knock on Alligator, I've seen it a million times. While it can be beneficial in improving the research, it also can be perceived as really jerky (esp. to people who aren't in academia). It is always easier to critique a study than to conduct one yourself. My apologies, Alligator, if my comment came across jerky.

      Anyway, none of that matters to anyone outside of grad students in quantitative programs, faculty who are obsessed with statistics and methodology and people who review/edit quantitative journal submissions (this last group is comprised of people who had previously been in the other groups).
    1. Alligator's Avatar
      Alligator -
      1.
      Quote Originally Posted by dje97001 View Post
      Alligator, the value is in that looking at hikers who finish the whole thing may provide lessons as far as pacing (esp. at the beginning when you are worried about people passing you, or wondering whether you started a week or two too late to make it to ME in time) and setting more realistic approximations for mail-drop locations.
      2.
      Quote Originally Posted by dje97001 View Post
      Yes, it is obvious this isn't a predictive model... but who cares?
      IMO, the above two statements are contradictory. But, referencing 1., that's why a confidence interval/s.e. would be useful. If it says that it takes a mean of 8 days +/- 0.5 days that would be helpful. If however, it says it takes 8 days +/- 3 days, that creates a different situation. See?

      Quote Originally Posted by dje97001 View Post
      It might be interesting to examine those who don't finish and compare daily mileage... maybe they pushed themselves too hard too fast.... maybe they realized that they would never make it to maine at the speed they were capable of... all of it is interesting stuff.
      Sure.
      Quote Originally Posted by dje97001 View Post
      Yes, it is obvious this isn't a predictive model... but who cares? Consider the years he examined to be the population that exists.
      It is most certainly a sample and not the population.
      Quote Originally Posted by dje97001 View Post
      The means for the population may be significantly different than the means for those who don't complete the thru, those who don't journal online at trailjournals.com, those who completed in a previous year (not included in the analysis), or those who completed the hike by sections, etc. ... but he didn't claim this extended to those pops.
      I fully understand that. But if there are actual differences, saying the mean is the same for all groups (even just thruhiker groups) masks what may be important underlying differences. Map man is making a serious effort. I understand that, that is why I gave what he presented a serious review. Items placed up for Articles are subject to review. It makes them better. It's extremely common to have things reviewed. Relax, I didn't mark the article rejected for review box .

      Quote Originally Posted by dje97001 View Post
      Still, to worry about kurtosis or skewness in this case is pretty much a waste of time--most people don't even bother checking for that stuff anyway... they just live and die by the central limit theorem. Sure, an exceptionally rainy year may have slowed people early on (resulting in more zero days) but then really dry years may have resulted in faster paces (with fewer zeros). It should all even out (life is all about probabilities).
      But if there are differences, that evening out may not have any meaning.
      Quote Originally Posted by dje97001 View Post
      We're only talking about 5 years (and ideally we'd have more) so there is likely to be a larger SE than you'd like, but without anything else to use, this is damn good stuff. You could always compare the 3 measures of central tendency to get a better idea of whether or not you have some outliers screwing with the data if you are really worried about it, but again, why bother, this is really interesting to chew on. Thanks map man!
      I didn't mention any distributional problems. Given his 105 samples, I wouldn't expect serious problems with his means. Also, I actually liked that he presented the medians as an alternative. He should be cognizant of any extreme outliers though. Further, I only asked for the SE, I haven't commented on how large it may or may not be.

      Sure, a lot of effort was put into it. It may certainly be reasonable to use. But care needs to be taken to ensure that any underlying biases are at least considered and hopefully controlled for. For instance, wouldn't it be interesting to know if there are age differences among hikers and if the sample was representative age-wise? Someone previously mentioned they pick a hiker similar to themself to compare to. And wasn't there a really wet year on the AT in that pool of data? Do Feb. starters really complete the trail on average the same as April hikers? It is important to consider factors such as these and not to immediately discount them.

      It is good stuff! I'd consider using it.
    1. dje97001's Avatar
      dje97001 -
      WRT the "contradiction"... I was making the statement that it wasn't predictive in the sense of an academic model. But there are massive differences between theory and practice. Theoretically, you can't make causal assumptions about correlational data (unless you've taken care of all of those pre-requisites, i.e. temporal ordering, etc.)... but practically? I would definitely use this data to "predict" where I will be at x days out... especially since this is the best compilation of numbers that I've seen.
    1. carolinahiker's Avatar
      carolinahiker -
      Im goin to section hike from erwin tenn to hot springs nc in may has anyone done the section lately and whats it like trail wise ? Thanks.

      Rick
    1. Alligator's Avatar
      Alligator -
      Quote Originally Posted by dje97001 View Post
      I get all of that. I think we all do. But you know as well as I do that very few studies actually use anything other than convenience samples (which are non-random--I'm not talking about random assignment here) because of cost, time and difficulty in compiling the true pop list. So basically we all assume normality, again unless we look at the skewness (or how flat or peaked the distribution is), which no one does. I'm sure you could do it, but again, I'm not sure what good it would do. Just compare the mean to the median... the closer they are to eachother the less likely skew exists.
      BTW, skewness refers to the distributions symmetry--heavy tails on the right, heavy tails on the left. Kurtosis refers to "peakedness".

      And I conduct my own research for applied science all the time.
    1. dje97001's Avatar
      dje97001 -
      Yeah thanks pal. I know that (lepto, platy, meso...). I was giving you options... consider it a grammatical mistake.

      The point is, whether you believe it or not, you and I probably agree on 98+% of this crap. Save possibly your stance on standardized vs. unstandardized (correlation vs. covariation) or maximum likelihood vs. least squares assumptions. I agree with your statements from a research standpoint. The funny thing is that those who haven't spent much time in stats/methods don't realize that these debates can be just as intense as the "hammock vs. tent" or even the dreaded "purist" debate.

      The point I'm trying to make--one that Chris made to me a while ago (but it took a while to accept)--is that there are already too many numbers for most people to spend much time on. Hikers understand that their mileage may vary. Confidence intervals, while very useful in gaining more precision (most of the time extremely desirable), in this case will simply obscure the value of these numbers to most people (i.e. they don't want to know that there is a 95% chance of them making it from Springer to the Georgia border in 7.25 to 8.45 days) they just want the best guess, for which the mean (or median) should suffice. Yes the pace for april starters may be different from march or feb starters for only the first 2 sections... but map man didn't ask for a critique. He didn't even ask that it be placed in the articles section. Such a detailed criticism without prompting will only make it less likely for people to share potentially valuable information. I for one, think that in its present form it is definitely of value to the hiking community.



      That set of numbers map man listed is complicated enough.
  • Recent RSS feeds

    WhiteBlaze

    Author Of Appalachian Trail Story To Hold Book Talk - CapeNews.net

    Author Of Appalachian Trail Story To Hold Book Talk  CapeNews.net

    More...

    WhiteBlaze 10-15-2019, 08:40 Go to last post
    WhiteBlaze

    Hiking the North Country Trail in Addison County - Addison County Independent

    Hiking the North Country Trail in Addison County  Addison County Independent

    More...

    WhiteBlaze 10-14-2019, 13:10 Go to last post
    WhiteBlaze

    Supreme Court showdown set over pipeline crossing of Appalachian Trail - Lynchburg Ne

    Supreme Court showdown set over pipeline crossing of Appalachian Trail  Lynchburg News and Advance

    More...

    WhiteBlaze 10-13-2019, 21:10 Go to last post
    WhiteBlaze

    'I needed to do it': Tabusintac woman conquers Appalachian Trail - CBC.ca

    'I needed to do it': Tabusintac woman conquers Appalachian Trail  CBC.ca

    More...

    WhiteBlaze 10-13-2019, 16:40 Go to last post