Register Forum Rules FAQ Members List Social Groups Search Today's Posts Mark Forums Read

oompa loompa
Senior Member

Join Date: Jun 2007
Location: 28° 37', North ; 77° 13', East
Age: 26
Quote:
 Originally Posted by Ledgem I don't disagree entirely, but why should we stop at 100 repetitions? That's nothing compared to 1,000 repetitions... or is it? Standard deviation isn't anything overly complicated, and there's no P-value. It's just a way of presenting some repetitions. For more complicated statistical analyses people generally present their number of data points (N). What constitutes a good sample size? That depends on what we're trying to observe, and we can judge the strength (or power, to speak statistics lingo) of the study based on N. But getting back to the repetitions of a single data point, something more than three points would be nice but I don't feel that it would be any more or less misleading. Seeing that something was done even three times convinces me that it wasn't just a one-time fluke, but I can still take issue with the technique or some other variable. Someone could repeat something even 1,000 times and it wouldn't answer that concern. This isn't to say that I'm against the idea of doing more than three repetitions, or even stating how many repetitions were performed.
Well yes, there is a reliable sample size, depending on the amount of variables, and the distribution it comes from, because you fix the probability of precision, (usually 95%, I don't remember if that's just convention or if it maximizes power somehow), so your gains from increasing sample size beyond a point decrease dramatically.

Also, since you calculate your SD from your variance, it is again important to have a large sample size because variance IS something you can test statistically, not with a p-value, but still. In this example, the number of variables isn't given, but if it's more than 3 I can definitely say that 3 is not enough. Besides, SD is a useless statistic without a sample mean (which I'm assuming is also being calculated), I'm a little rusty (and generally stats is not my strong suit) but since the sample mean is derived from a normal distribution by the CLT, there is going to be an associated t-statistic and p-value.

It's not misleading if you're not trying to mislead anyone. It is misleading if the claim is that an SD with 3 data points (and less than 3 variables) is a good predictor for what the deviations of the results of future experiments will be. For example, if you get a result greater than x number of SD's for your 4th experiment, how will you know if something went wrong or right? Forget SD, if you don't have a reliable mean, how will you know whether your result was close to what the 'usual' result is or not? If you don't have a reliable variance, how will you know how much deviation is acceptable without being called an abnormality?

At the same time, I agree with you that 3 is better than nothing. Its always going to help, but one would have to say that the results are merely indicative, and not a reliable predictor for how future experiments will pan out.

Last edited by oompa loompa; 2014-05-15 at 04:22. Reason: wow.. i need to sleep before i post

 2014-05-15, 12:28 Link #7502 Frailty God General Graphic Designer     Join Date: Jan 2008 Location: Pearl of the Orient Seas Age: 23 Sometimes, whenever I upload a jpeg image on the internet, the image quality somehow downgrades. Any particular reason for this? Is it the bandwidth? how can I upload the image with assurance that the quality won't be lowered? sorry for the multiple questions __________________
 2014-05-15, 15:03 Link #7503 bhl88 Otaku Apprentice     Join Date: Jun 2008 Location: The Unseen Horizon Is there a proper way to use a bow? (as far as I know, they say that she's handing the bow wrong, but I don't know why [some ads NSFW]) __________________ Dang it Avalon, you c(XD LOL)-block Shirou and Reinforce, but don't protect his mind in other ways? What is wrong, you woman? Friendship, be made! Magical power, gather! Starlight Breaker.... this world!
Dextro
He Without a Title

Join Date: Feb 2008
Location: Lisbon, Portugal
Age: 29
Quote:
 Originally Posted by Frailty Sometimes, whenever I upload a jpeg image on the internet, the image quality somehow downgrades. Any particular reason for this? Is it the bandwidth? how can I upload the image with assurance that the quality won't be lowered? sorry for the multiple questions
I don't think there's anything you can do. Some websites re-encode the jpeg you upload which inevitably lowers quality. Jpeg is a rather complex format that doesn't really store colors but frequencies, I believe, which makes it very adversely affected by re-encodings (just cropping a jpeg can lead to loss in quality). To top it off plenty of sites force the quality settings of the algorithm to a very low size setting which leads to even more compression artefacts. The users really can't do much to fix this.
__________________

Kotohono
Yuri µ'serator
Moderator

Join Date: Nov 2009
Location: FL, USA
Age: 28
Quote:
 Originally Posted by Frailty Sometimes, whenever I upload a jpeg image on the internet, the image quality somehow downgrades. Any particular reason for this? Is it the bandwidth? how can I upload the image with assurance that the quality won't be lowered? sorry for the multiple questions
What website are you using to host as different image hosts have different limitations on how and what you can upload.
__________________

Sig by Patchy
Avatar by TheEroKing
MAL

Frailty
God General
Graphic Designer

Join Date: Jan 2008
Location: Pearl of the Orient Seas
Age: 23
Quote:
 Originally Posted by Dextro I don't think there's anything you can do. Some websites re-encode the jpeg you upload which inevitably lowers quality. Jpeg is a rather complex format that doesn't really store colors but frequencies, I believe, which makes it very adversely affected by re-encodings (just cropping a jpeg can lead to loss in quality). To top it off plenty of sites force the quality settings of the algorithm to a very low size setting which leads to even more compression artefacts. The users really can't do much to fix this.
Well that sucks. The png versions of my tags aren't usable here.

Quote:
 Originally Posted by Konakaga What website are you using to host as different image hosts have different limitations on how and what you can upload.
I'm using imgur. I've been using that for quite a while but it has never happened to me before in imgur
in fact, I switched to imgur from photobucket because of the quality degrading thing.

thank you guys
__________________

 2014-05-16, 02:27 Link #7507 Dextro He Without a Title     Join Date: Feb 2008 Location: Lisbon, Portugal Age: 29 You can always upload the png file to imgur instead of a jpeg. Png doesn't have to the same scaling issues as the later format so it might help. __________________
risingstar3110
✘˵╹◡╹˶✘

Join Date: Nov 2006
Location: Australia
Quote:
 Originally Posted by oompa loompa Well yes, there is a reliable sample size, depending on the amount of variables, and the distribution it comes from, because you fix the probability of precision, (usually 95%, I don't remember if that's just convention or if it maximizes power somehow), so your gains from increasing sample size beyond a point decrease dramatically. Also, since you calculate your SD from your variance, it is again important to have a large sample size because variance IS something you can test statistically, not with a p-value, but still. In this example, the number of variables isn't given, but if it's more than 3 I can definitely say that 3 is not enough. Besides, SD is a useless statistic without a sample mean (which I'm assuming is also being calculated), I'm a little rusty (and generally stats is not my strong suit) but since the sample mean is derived from a normal distribution by the CLT, there is going to be an associated t-statistic and p-value. It's not misleading if you're not trying to mislead anyone. It is misleading if the claim is that an SD with 3 data points (and less than 3 variables) is a good predictor for what the deviations of the results of future experiments will be. For example, if you get a result greater than x number of SD's for your 4th experiment, how will you know if something went wrong or right? Forget SD, if you don't have a reliable mean, how will you know whether your result was close to what the 'usual' result is or not? If you don't have a reliable variance, how will you know how much deviation is acceptable without being called an abnormality? At the same time, I agree with you that 3 is better than nothing. Its always going to help, but one would have to say that the results are merely indicative, and not a reliable predictor for how future experiments will pan out.
Sorry guys were, out of it till now.

Of course in an ideal world, we will like to set out a survey of 100 samples with 100 duplicates, but general we don't have enough time and human resources to follow up with that.

What I did was basically set up an experiment which from 10 data point, to produce a trend (something linear, like interviewing 10 people to link the coffee consumption with ages for example). I generally agree that a repeat have to be done to make sure that the results will be accurate.

But when it comes to data analysis, what i think we have to do is: repeat the experiment, and then set up a trend based on the 20 data points (10+10), and draw a line on that, taking the R^2 as variance

My boss meanwhile want to repeat the experiment three times (interview 10 people each), draw a trend, get the slope each time, then do standard deviation(SD) for those 3 just to get SD, which i found really absurd. I means I saw guys in my office even try to put in that +/- sign for his data using only 2 data points before, just to suit my boss taste.
__________________

oompa loompa
Senior Member

Join Date: Jun 2007
Location: 28° 37', North ; 77° 13', East
Age: 26
Quote:
 Originally Posted by risingstar3110 Sorry guys were, out of it till now. Of course in an ideal world, we will like to set out a survey of 100 samples with 100 duplicates, but general we don't have enough time and human resources to follow up with that. What I did was basically set up an experiment which from 10 data point, to produce a trend (something linear, like interviewing 10 people to link the coffee consumption with ages for example). I generally agree that a repeat have to be done to make sure that the results will be accurate. But when it comes to data analysis, what i think we have to do is: repeat the experiment, and then set up a trend based on the 20 data points (10+10), and draw a line on that, taking the R^2 as variance My boss meanwhile want to repeat the experiment three times (interview 10 people each), draw a trend, get the slope each time, then do standard deviation(SD) for those 3 just to get SD, which i found really absurd. I means I saw guys in my office even try to put in that +/- sign for his data using only 2 data points before, just to suit my boss taste.
So you have 30 data points. Don't analyze each trend separately, you can assume that each data point is drawn from the same population distribution. If the sampling is done randomly, just assume that they're all from one sample. 30 data points is not much, but its not bad either. You're going to want to run an OLS regression, and you'll probably want to control for other things apart from age, but I leave that up to you.

If you can, try to get a statistical software, either R or Stata ( I think R is free, but stata is hella expensive, you can download it illegally if you want), and they will do all the analysis for you. You could also do it on excel, but its a little more complicated. At any rate I don't think you want to be calculating this stuff manually. Also R^2 and variance are not the same thing, A simple way to look at R^2 is the amount (%) of variance explained by your regression. I really recommend you do it with some statistical software, it will make your life soooo much easier. Just put your data in, write one line of code, and poof! you'll get a full statistical summary with everything you need. It will also summarize your raw data for you, with your means, SD variance, etc.

 2014-05-20, 07:25 Link #7510 Haak Senior Member     Join Date: Apr 2009 Location: UK Age: 25 Do the phrases "fuck my life" and "YOLO" have Japanese equivalents? __________________
Dextro
He Without a Title

Join Date: Feb 2008
Location: Lisbon, Portugal
Age: 29
Quote:
 Originally Posted by Haak Do the phrases "fuck my life" and "YOLO" have Japanese equivalents?
Does "Fukou da" work for the first one?
__________________

 2014-05-21, 07:13 Link #7512 Haak Senior Member     Join Date: Apr 2009 Location: UK Age: 25 That's more of a less obscene version of "Fuck me" than "Fuck my life" but I guess it would have to do. __________________
 2014-05-28, 10:28 Link #7513 risingstar3110 ✘˵╹◡╹˶✘     Join Date: Nov 2006 Location: Australia Anyone think too many half-Japanese character in recent anime/manga/ LN had dad as Japanese and Mum (or sometime grandmother) as foreigner instead of another way around (despite the other is much more common)? Was it something negative to male audience in general (as Japanese girls choosing a foreign man), so the author have to avoid backslash? If not, do you know many recent anime with half Japanese characters having foreign dad? __________________
 2014-05-30, 02:36 Link #7514 Irenicus Le fou, c'est moi     Join Date: Dec 2007 Location: Las Vegas, NV, USA Age: 27 I get the feeling they sort of interpreted "dominant" ethnicity based on the father side, so an offspring of a Japanese father x foreign mother is a Japanese with exotic brushes, while the other way around is actually an exotic foreigner who has some Japanese ties. So it's like, "how come you can speak English/French/Sindarin so well?" "Well, my mom's elvish," vs. "wow, your Japanese is good!" "Thanks, my mom is Japanese, so..." *Speaking only for anime/fiction. No idea how halfs are interpreted socially in reality.
 2014-06-07, 12:20 Link #7515 Tiberium Wolf Senior Member     Join Date: Dec 2004 Location: Portugal Age: 36 What's the difference between LN and Web novel? I see translations of LN but no web novels. __________________
 2014-06-08, 15:43 Link #7516 risingstar3110 ✘˵╹◡╹˶✘     Join Date: Nov 2006 Location: Australia This could belong to the old silly thread: But you know how sometime people suggested you to "go see your doctor" for annual check up or things like that, who exactly are these doctors? This could be silly to say, but I never got sick before. I still remember the last time i got a cold (the normally bed-ridden kind in anime), and it was 10 years ago, but I even dragged myself to school and had my drama class in the freezing hall. Never break a bone, strain my muscle, or anything that requires more than an hour of sleep to recover. That's why when someone suggested me to "go see your doctor for annual check up" it just puzzle me deeply. Who is my "doctor" in this case? My best guess is a private doctor who I should have seen form time to time, and who under covered by the state universal healthcare (called Medicare here)? How can I pick up a doctor in this case? Just look for nearby clinic? Or do I have to go to a hospital for it? __________________
 2014-06-08, 16:24 Link #7517 monster Junior Member     Join Date: Dec 2005 I don't know how it works in Australia, but in the US, it's usually your general practice/family doctor that you get from your insurance company, either picked by your insurance company or chosen by you amongst a lists of doctors that are contracted with your insurance company. I suggest either to ask your insurance company for a doctor or, if that's not how it works in Australia, then maybe you can go to your local hospital and ask about information on how to get a doctor.
 2014-06-09, 22:44 Link #7518 risingstar3110 ✘˵╹◡╹˶✘     Join Date: Nov 2006 Location: Australia More question if everyone don't mind Haven't made a ppt since uni time. But back then we didn't have to worry about the whole copy right issue. Now can't just google image and copy down whatever i see fit anymore. Hence was just wondering if there is any website that lists open-source (free to use) pictures or illustrations that we can use freely on ppt without any concerns? The list in Microsoft PowerPoint is too limited for me __________________
 2014-06-09, 23:24 Link #7519 erneiz_hyde 俺強えええ (笑）     Join Date: Sep 2010 Location: InterWebs It's just a presentation right? Can't you just cite the source you take it from? Like below the picture or in the very back? Didn't you have to do that in uni as well? Unless you plan on using said images in commercial products like magazines, printed shirts, etc I don't think there should be a problem, though cmiiw. __________________
risingstar3110
✘˵╹◡╹˶✘

Join Date: Nov 2006
Location: Australia
Quote:
 Originally Posted by erneiz_hyde It's just a presentation right? Can't you just cite the source you take it from? Like below the picture or in the very back? Didn't you have to do that in uni as well? Unless you plan on using said images in commercial products like magazines, printed shirts, etc I don't think there should be a problem, though cmiiw.
Really?

I'm not really knowledge on those copy right law and issues. Thought i have to send email to every website or blog, asking whether i can use their images

Found a decent one here as well: http://www.freedigitalphotos.net/
__________________

 Tags problem, q&a, questions, serious