Magic Numbers: Downloads > Retail

By Alec Meer on September 20th, 2010 at 10:05 pm.

So says US market research group the NPD, who currently rule the roost in terms of game sales figures by dint of no-one else doing it properly. Today, they give with one hand and take with the other.

Give! US PC game download sales for the first six months of the year popularly known as Twenty Ten finally outweigh old-fashioned high street retail – by 11.2 million to 8.2 million units, to be precise. Well, I say precise… we’ll get to that in a minute.

Take! PC sales as a whole are lower than at the same time last year. 14% lower, to be precise. Well, I say precise. We’ll get to that in a minute. Actually, how about right now? What I really mean is “at a guess, because this isn’t actually based on sales figures and thus may well be utterly meaningless.”

Here’s what the NPD claim are the biggest download sites for the first half of this year (only for the US, remember):

Top 5 Frontline Digital Retailers –Jan.-June 2010 (based on unit % share)

1. Steamgames.com
2. Direct2drive.com
3. EA.com
4. Worldofwarcraft.com
5. Blizzard.com

Which is a whole fat bunch of interesting. Except… in my various other escapades, I’ve been told by the heavyweight likes of Valve and Blizzard that, despite being kings of these charts, they do not provide sales figures to the research group. Or, indeed, to anybody. The same very much appears to be true of every other download service – which makes where that 11.2 million sales figure came from a little opaque.

The answer to this conundrum is relatively simple: the NPD survey a bunch of gamers, multiply their answers several hundredfold, average it out, make a few educated “projections” and bingo. Exactly 11.2 million sales. Neat and tidy and half the internet blindly reports it as gospel truth because it’s a number and it’s come from a market research firm therefore it must be right.

Small print is still print. Read it, think about it, question what’s in the big print, by all means base your sense of what’s going on around the information there within, but don’t take it as an exact science and the final word. Please. Enough of this kind of reportage.

I’m not taking a pop at the NPD here. They’re clearly trying exceptionally hard, even if they should have started doing this years earlier. If the sales figures aren’t made available, the sales figures genuinely aren’t available – so they’ve gone to great lengths to try and read the lie of the land and thus this is genuinely as good as we’re going to get. It’s useful to have any information whatsoever on what’s going on in Mysterious Digital Land, and worth being grateful that someone’s trying to do that for us. But that most of the world’s games journalists and interested bystanders are currently basing their picture of PC Gaming in 2010 almost entirely on this set of survey-derived numbers is a little unsettling.

Even more so because of the omissions. Does that cover FarmVille? Minecraft? Love? D&D Online’s Free To Play relaunch? GoG’s retro sales (speaking of which, they’ve posted an update, which seems to somewhat corroborate mounting speculation that they’re not even slightly dead)? A hundred, a thousand, a hundred other thousand weird’n'wonderful’n'copycat
‘n’cynical’n'awful’n'amazing titles that lurk out there, stealing people in handfuls or barrow-loads. Are they represented in that 11.2 million and that -14%?

No, probably not. And thus to cheerfully wipe 7/50 health off PC gaming based on what scant information is available just seems ludicrous. For all I know it’s declined far more than that. But I suspect the sum total is in fact on the up, that these figures paint an image of an ailing platform when in fact it’s quite the opposite. Yeah, PC Gaming in its traditionally-perceived form, the kind of thing that we’d post videos of from Gametrailers, is perhaps not experiencing its rudest health right now. But PC gaming, the whole great, infinite mass of it? C’mon. You need more than a survey a few thousand people who fill in forms online to measure that ever-changing beast’s flabby, beautiful bulk.

Here’s that small print explaining the NPD’s methodology, by the way. Decipher!

- Games Acquisition Monitor

The NPD Group’s Games Acquisition Monitor is a quarterly tracker which measures both digital and physical forms of games acquisition activity, volume, awareness and usage of retailers and services, as well as other technology and entertainment activities that could influence game acquisition trends. It is based on online survey responses from over 8,000 members of NPD’s online consumer panel. Data is weighted and projected to be representative of the U.S. population ages 2 and older.

- NPD Consumer Tracker

NPD collects data on PC game purchases via its weekly video games consumer survey. Each week, over 180,000 individuals are selected from the NPD online consumer panel to participate in one of four weekly studies. The responding sample is demographically weighted and projected through a series of steps to represent the Total Adult (18+) and Total Teen (13-17) U.S. population. Respondents to the survey report whether in the past week they purchased a PC game on a disc or downloaded the game from a website directly to their computers.

Only digital purchases of games from the above websites are tracked in this report. Shipped boxed products from these sites are not covered; information about these can be obtained through our PC Retail Tracking service.

Hmm. 8000? Or 180000? And how do any of these people end up on the “online consumer panel?” Again, I’m not arguing that these figures aren’t reflective of the market, that NPD have done anything wrong, and most certainly not that anyone’s doing it better. But it’s important to only assess and digest the information they give us in the context of how it was gathered, and not just see site after site report the headline numbers as Ultimate Truth. Downloadable and online gaming is huge in both stature and moneyhats, it’s growing constantly and it’s almost impossible to comprehensively document. We can’t pretend to have its entire measure from survey results alone.

__________________

« | »

, .

73 Comments »

  1. Andreas says:

    I think you’re underestimating how correct that 11.2 million figure might be – statistics, as much as people hate to admit it, is a pretty damnably exact science, as ridiculous as that statement may sound.

    I’m not saying we should take their figures at face value, but far, far too many people automatically assume a sampled survey means bugger all.

    • Starky says:

      Also, Steam/Valve may not release sales numbers, but EA might release their steam sales to the NPD, likewise with Activision and so on and so forth – which will probably give them a good idea.

      Say if the top 10 games they have solid number from the publishers sold 20% more on DD compared to physical over all platforms, and 70% of those DD sales were on steam – you could start to build a pretty solid set of statistics and assumptions.

    • DJ Phantoon says:

      But people also accept statistics at face value. It’s more dangerous to do that than to remain skeptical because math is only so accurate as the person who inputs values. Human error is more likely than perfection in prediction.

    • Archonsod says:

      Surveys are problematic at generating any meaningful data in the first place. When you start putting them up online so you can’t even exercise minimum control over the sample cohort you can pretty much give up any hope of obtaining accurate data. There’s as much chance the survey reflects the attitudes of a guy in Uzbekistan with a proxy and a desire for an Ipod as it does anything you’re trying to measure.

      Plus it’s the worst way to measure this. A better method would have been to take something like SC2, try and get hold of the figures for online vs retail sales and then extrapolate that out as a percentage. You’d risk missing the more niche results (I imagine for example 99.9% of modern indie games are sold digital only) but it would give you a general indicator of the market, and then of course you can do the same thing with other major releases and chart the difference.

    • battles_atlas says:

      Absolutely on the contrary Andreas, clearly far too many people see a statistic and simply take it at face value. Certainly the media do, and report it as such. Those readers who are a) otherwise informed; and/or b) find the results conflict with their ideology; will likely reject or at least question the numbers, but everyone else will likely file them under “Fact”. Given that the perception promoted by the media is already along the lines that PC gaming is dying, results like this probably aren’t going to be questioned to any great degree by most who read them.

      Whilst we’re on the subject, statistics – by which I mean their misuse – are the bane of modernity. For a start our entire economic system is based on metrics which fail to account for externalities and thus give no real clue of profit or loss. Education and health is weighed down by mountains of poorly considered metrics. GDP has become the overriding measure of human achivement, at great cost (ironically), simply because actually useful measures like social cohesion or individual happiness defy easy reification. The unfortunate truth is that there is much in the world that can’t be captured, at least with methods available to us, in a handy numerical form. I hope the power brokers figure that out at some point.

      In other news, I’m feeling much emboldened today regarding the health of digital entertainment, having just discovered the fantastic music site http://www.bandcamp.com. At last an outlet that really can free artists from the clutches of the Suits. Wonderful case in point: http://cloudkicker.bandcamp.com/album/beacons.

    • TheApologist says:

      @battle_atlas
      Very well put – hear hear

    • Rinox says:

      @ Battle_Atlas

      re: GDP. This is true. Recent research for example showed that a nation’s general health was not tied quite as much to its GDP but to its ‘equality’, i.e. the distribution of wealth within the country. This is hardly surprising, but not something that is considered by many observers. A country may have a great GDP but if only 10% of your population only really profit from it…then yeah it’s probably not that much of a country you want to live in.

  2. Zyrxil says:

    I’ve always wondered how they determine TV ratings. I’ve read about “Nielsen families”, but that can’t be the end of it, right?

    • bob_d says:

      It’s crazy, but yeah, that’s pretty much what the entire US television industry runs on, despite it being acknowledged that the numbers aren’t all that accurate in that case (in part because it isn’t a completely random sample).

    • ScubaMonster says:

      Wow that’s insane. I figured they had a way to cull the data from cable and satellite providers. Kind of sad and scary at the same time that our favorite shows live or die based on “Nielsen Families”.

    • bob_d says:

      You listen to industry people talk, and they say, “Yeah, the system’s broken… but it’s the only system we have, so we have to use it.”

    • Andrew Dunn says:

      It’s a similar situation in the UK. Certain households are tracked for ratings and the figures are extrapolated from there. Pretty worthless figures, really.

    • TNO says:

      Actually, the industry guys know it’s broken, and the KNOW they could do better in today’s world by actually digitally tracking people’s behaviour (with consent, of course). the problem isn’t technical, it’s psychological.

      The TV industry has decades of practice built around the currently, flawed system. Sweeps week? Yup, that’s designed to boost Nielsen numbers. The rates advertisers pay? Built around Nielsen numbers.

      If they were to switch over, every TV network is scared about what they could learn. Maybe CSI actually isn’t the number 1 show? Perhaps people who watch American Idol are more likely to change channels during commercials? Suppose only 3 people actually watch Spike?

      Because the stakes are so high and involve so much money, everyone is too scared to change the way the data is gathered because they are too afraid of shattering the business model. I’m not knocking them for this, I understand the reasons behind it. It’s basically saying “better the devil we know…”

    • Xercies says:

      Of course its utterly broken and is pretty much dead in the way people watch things now a days, online and all that. It realy pains me to see some of my favourite shows being cancelled because of an antiquitated system but theres nthing really we can do about it that much. Business will use the system until it crashes on them.

  3. LewieP says:

    Consoles are pretty popular in the USA too. PC gaming is no doubt on the up in the places that Nintendo, Sony and Microsoft don’t sell their consoles.

    Also, it’d be dangerous to extrapolate USA trends globally.

    • subedii says:

      Pretty much. PC games sales tend to be stronger in Europe compared to the US to begin with. Then you’ve got a lot of East Asia that’s highly developed and where PC gaming has a heavy presence. Granted a lot of that is things like F2P games, but those still probably rake in a LOT of money.

  4. Po0py says:

    Can I just say that PC gaming if doing fine and we needn’t worry about it too much.

    There. Said it.

  5. Kurina says:

    I have no doubt the staff of NPD is capable of gathering information and giving fairly relative statistics. My problem is the sample pool they use. These are not people answering surveys out of the goodness of their hearts, and trying to offer information to help corporations. A majority of people involved in these surveys are doing so because prizes are involved.

    Technically that should be no problem, but many people sign up for surveys and do not bother answering them truthfully. They click random answers, so they can get to the prize. All you have to do is browse forums dedicated to survey groups, and see people discussing the best way to obtain credit and prizes by the answers they choose.

    I would be more accepting of these results if the sample groups were handled differently. Just my two cents anyways.

  6. Gamerdude says:

    Statistics DONE RIGHT can be very accurate. When measuring card or dice odds for example.

    Applying these as blanket demographics is far less accurate.
    How these 8000 or 180000 people for a survey are chosen is an important factor.

    • Thermal Ions says:

      Not to mention, how many of these selected members actually complete the survey in that week? Are they actually calculating off a % of the 180,000 who respond or taking the first 180,000 respondents from a larger pool of invites they send out?

      And when it comes to this statement: “as well as other technology and entertainment activities that could influence game acquisition trends” – there’s a whole load of grey assumptions going on.

  7. Nemon says:

    US market trends? WHOAHw the science!

  8. StingingVelvet says:

    Three kinds of lies… lies, damn lies and statistics.

    I told my statistics professor in college that and kind of shrugged and said “yeah.” It’s not that they’re really lies of course, but more… assumptions. That same professor compared statistics to assuming a hot young blonde married to an old rich man is a gold digger. You play the odds, based on social trends and established stereotypes.

    In any case, I still buy boxes, but I don’t NEED boxes, which is to say I buy download only titles as well. If one day everything is download-only it’s not like I will stop buying games.

    Streaming-only yes, download-only no.

    • Poindexter says:

      This was my statistics professor’s view on statistics: “Statistics are like a bikini; they certainly reveal a lot, but not what you really want to see.”

  9. Matt W says:

    Second Andreas – from what I understand, serious surveying companies are fairly scarily good at their meat-and-potatoes work (which consumer spending almost certainly qualifies as). Yes, it’s unlikely that their base sample is not representative of the market as a whole, but I’m fairly confident that they’ve thought of that and are correcting for it.

    It is entirely possible (likely, even) that their methodology has kinks, and it’s possible that those kinks are throwing the numbers unacceptably out of whack, but on balance I’d be tempted to go along with them until such time as someone who actually understands the sorts of calculations involved suggests we should do otherwise. Automatically casting aspersions at the validity of any result we dislike gives an impression of churlishness – ask yourself, if the news was all positive, would you still publically question the methodology?

  10. Starky says:

    I’d be willing to bet these statistics are vastly more reliable than 99% of make-up, bathroom, or grooming products.

    BEST SHAVER EVER! (and in small writing at the bottom “70% of 68 people agree”)… so on and so forth.

    Hell give me a free Gillette 9 blade hyper shaver 10 and I’ll agree to whatever you like in a survey.

  11. Skippy says:

    As someone who rages daily about shitty polling, thank you. This is my favourite RPS post ever.

  12. Jimbo says:

    Do Impulse and GamersGate still demand respect?

    The far bigger issue with these stats is that that they still don’t have $ signs next to them. Selling more units is pretty irrelevant if they’re all being sold for $3 in a Steam sale.

    • DrGonzo says:

      I think what’s more important is profit rather than how much it’s sold at. GTA sells by the bucket load but seems to struggle to make a profit.

    • Wolfox says:

      Gamersgate, yes. Impulse, not really.

    • Jimbo says:

      Sure, but realistically the NPD can’t comment on profitability. They ought to be able to comment on the dollar value of the retail sales and digi sales though, at least as accurately as their survey allows. And if they can’t do that, then the survey doesn’t really tell us anything useful.

      For what it’s worth, US console software $ sales in August were down 14% on last August, so it’s not like it’s just the PC market taking a kicking at the moment.

      It’s also worth noting that these PC figures don’t include StarCraft 2…

    • DrGonzo says:

      So essentially all game sales have dropped by 14%?

    • pkt-zer0 says:

      Heh, SC2. Some more numbers from NPD: sales in July were down 15% compared to last year on consoles. PC games were up 103%, largely attributed to three days of Starcraft 2.

    • Jimbo says:

      If you look at two particularly gloomy pieces of data, you could roughly say that, ish (they aren’t exactly like for like).

      I don’t think they’re a particularly accurate reflection of how the market really is though. It’s bad but not that bad. I suspect if you compare the whole of ’10 to the whole of ’09 then game sales will be down, but nothing like 14% down. August was just a particularly bad month for console releases this year, and the PC data doesn’t include StarCraft 2 or Civ 5 data yet, which out to pull the PC numbers back up a bit towards the end of the year.

  13. Chuckchuckrazool says:

    I’ve completed several advanced statistics courses in college and graduate school (so I’m no expert), and though I agree that as tools they can be remarkably accurate, it looks like this sampling methodology has a number of weaknesses. Arriving at accurate estimations of these numbers on a global scale, using these methods, is probably not possible. We might get pretty good estimations of the relative proportions of survey volunteers’ game-buying habits (i.e., 20% use Steam, etc.), but yeah, extrapolating the sample into global figures for the entire gamer population is a bit sketchy. I think you nailed it in the article, Alec: people like any print more than no print, and if I’ve learned one thing in stats, it’s that bad figures can be worse than no figures.

  14. teo says:

    Why do we as gamers have to worry about the industry so damned much? The day PC gaming is dead I’m sure we’ll notice

    • Bob Bobson says:

      Have a bonus point. And a cookie.

      If publishers think gaming is dead on the PC they won’t publish PC games. But publishers know how their own sales are going and can make decisions based on much better data than NPD have. Indies will choose the PC for their game development cause consoles are so locked up.

      There may be up turns and down turns but the only way I’ll know if PC gaming is dead or not is if I can buy and play PC games or not.

    • subedii says:

      I dunno, how can we be sure it’s not dead?

      Martha, fetch me my poking stick

      *poke*poke*…

    • bob_d says:

      PC gaming won’t ever “die” entirely; that’s not going to happen. But it can get very, very sickly. What you’d see is fewer and fewer profitable game companies, fewer big games (and the big games will also be on consoles), less risk-taking with game design (so fewer varieties of games, more sequels and more formulaic games), a lot of money being put into very simple, mostly identical web games, and more and more ridiculous anti-piracy measures being put into place.

      What we’re seeing happening in the industry now, essentially.

    • subedii says:

      Have you got any numbers to show a year on year decline since 2001? I’d be interested in seeing that.

      Also, are we talking global here or what?

    • Xercies says:

      The thing is unfortunatly major publishers see this data and refuse to make games on the PC because of the percieved nation. One of the reasons Epic did the very same thing was because of NPD figures. now say what you will about Epic games, that means were actually not getting those games AT ALL. And i have to say thats definitly bad, probably Microsoft also did this and maybe many other companies do this. So these figures, no matter how rubbish they are do affect PC gameing quite a bit.

      And that worries me.

  15. SorenGrey says:

    This is a fantastic example of theoretical thinking in games journalism! To read the fine print and really talk about what an article *means* — not just mindlessly regurgitate the information within it — is especially important in cases like this where data isn’t forthcoming. In the absence of hard statistics from developers, the health of the PC gaming field is constructed on such articles, and that means a great deal for the hobby we love because such constructions can go on to influence (or perhaps become) reality. Good deconstructions like this one can provide a check to that process by helping us take a step back and really look at what’s going on before we leap to judgment. In the process, we get a more accurate picture of the state of PC gaming, and developers and consumers can take more confident steps based on that picture, whether good or bad. Thanks for helping preserve a hobby we love.

  16. Jeremy says:

    As long as I own a PC, it won’t be dead.

  17. Wooly says:

    Anyone else notice that GOG has a new logo/symbol for the website? It seems unlikely to me that a service going out of business would give itself a new logo to commemorate it’s demise…

  18. Rich says:

    4. Worldofwarcraft.com
    5. Blizzard.com

    Up until StarCraft 2, wouldn’t WoW be the only thing one might buy from either site?
    Also, are we to assume they haven’t included buying game time for WoW?

    • DrGonzo says:

      If they included that PC gaming would dwarf all other platforms.

    • bob_d says:

      Yeah, and when two of the top 5 download vendors are representing the same game, that tends to indicate you don’t have healthy a download sales dynamic. (Or it indicates the numbers are way off, which I actually don’t think is the case here.)

  19. Nick says:

    Awesome picture.

    When attractive ladies didn’t have to be skeletal thin.

  20. bob_d says:

    When development costs are going up, it doesn’t really matter if sales are steady or going down – if the sales numbers aren’t going up, it’s bad for the industry. (Same thing is true with consoles, but they are demonstrably experiencing a sales slump, which makes the trend easier to see.)

  21. lostrootpass says:

    The whole “digital outselling physical” thing is a self-fulfilling prophecy, though. There will always be those like myself who prefer physical, boxed copies of games but we’re in the minority, apparently. Then when you consider that retailers like Game have went console-only and removed their PC section from all their Game & Gamestation stores (aside from one obligatory, barely noticable shelf tucked away in the corner), we have nowhere else to turn, much to our own dismay.

    Personally I think we should all just stop worrying about these things. PC gaming has been “dying” for 20 years.

    • DrGonzo says:

      Play.com or Amazon? Game etc may not be selling them anymore but it’s not really much of a loss, the internet is always far cheaper anyway and I imagine that had as much of an impact if not more than Digital Downloads.

    • Vinraith says:

      In the US, Amazon and GoGamer seem to be best for providing well priced boxed copies of PC games. I still prefer boxes myself, a position that’s become all the more entrenched in the wake of this GOG business.

    • subedii says:

      It would be more entrenched with me as well, except these days more and more games require online activation regardless of whether the game is store bought or downloaded.

      At which point the whole thing becomes moot pretty much, and I may as well just get it wherever it’s cheaper.

    • bob_d says:

      I assume that 20 year timespan wasn’t actually referring to anything real. There have been lots of ups and downs in the PC game business, but a serious decline started about 2001, with the release of the Xbox. Before that point, PC game sales were getting stronger each year.

  22. Novotny says:

    Surely Civ V sales will tell us all we need to know about pc games?

  23. Navagon says:

    It doesn’t seem to me like many people have a very comprehensive idea of what the PC market is like these days. I know I certainly don’t.

    At least we seem to have moved beyond thinking that the aspect of the PC market that is directly comparable to the console market = the entire PC market. But still, that realisation should have lead to the abandonment of statistics gathering rather than this wild guesswork.

  24. Cinnamon says:

    8000 respondents sounds like a pretty big sample.

    I’m tempted to believe this due to a lack of big releases and people being short of cash. What major PC releases were their in the first half of the year? Mass Effect 2 and Bioshock 2 come to mind but I don’t think that they exactly broke records on consoles. Would like to see how console game sales holding up in the same period but I don’t think there has been growth there, especially for Nintendo. Things like Starcraft 2 and Civ 5 might make the next report worth reading if they do it again.

    • Zyrxil says:

      One of the first things you learn when you study statistics is that sample size is not important beyond a certainly calculated threshold. Obtaining a sample that is representative of the population is far more important.

    • Cinnamon says:

      I’m not aware of what methods they used to verify that their samples were a good representation of the population but, you know what, I’ll just assume that they know what they are doing better than me or people who know even less than me but are just sceptical because they don’t like the results. With 8000 respondents they might do things like have multiple groups of samples that they compare to each other in different ways to see if they are consistent increasing their confidence in the results. They can never be 100% confident that their figures are correct but they have done more work that someone saying that they think that more people are buying PC games this year because PC games are great aren’t they?

  25. Bluebreaker says:

    PC gaming is dead ;-)

  26. Kevbo says:

    McCarthyism by numbers… I recently heard an interesting article on the radio about how marketers use numbers to deceive people. For example, when stating that something has a certain number of features or functions, don’t use a well rounded number like 10, 20 or 50. Instead use something like 41 or 23 or 108 since that tells humans that there is something concrete about that and that it is fact and not a generalization. I do agree that far too many people believe what they hear without any research and pass it along like its the secret truth that they have been let in on. Good to see some resistance and long live the king (pc gaming) :)

  27. pupsikaso says:

    Oh my gosh, what’s with that hair? I mean it’s bigger than her head! I mean… Yeah! Go downloads!

  28. Capon says:

    In a situation where a minority of survey forms deviate from the average, experienced statisticians will throw out any ‘done-quick’ results, like forms that were obviously clicked through to get the cake. They’d also throw out any forms that had obvious conflicting data errors:

    Q: How any hours a week do you play PC games?
    A: 10 hours

    Q: How many computer games do you own?
    A: Zero

    • Disquete says:

      That example may not be a conflicting error. In fact it is what a Farmville (or any other Facebook game) player would reply, unless the survey previously stated that those games do not count as PC games.

  29. The Sombrero Kid says:

    NPD are plain and simply trying to spin the truth into something desirable, the truth being that they are loosing their grip on their monopoly and that represents their ultimate demise, i suspect there is still value in counting and selling game sale numbers and so the digital stores will inevitably do it, but the NPD knows it wont be the one collating the data.

  30. Dreamhacker says:

    So, the sum of this article is: NPD is making crap up because they don’t actually have any real data on digital sales but still want to sell statistics?

  31. Tei says:

    What we need is a study of the PC population.

    1) The nintendo generation. How much nintendo guys have joined the PC numbers?

    2) The rise of the FPS on consoles. How much “core gamers” we have lost to XBox and the PS3?

    3) The deep pit of WoW. How much PC players are lost for gaming, and only, *only*, play WoW.

    4) The rise of the casuals. How much more gamers the PC has win thanks to casuals games, webgames and indie games?

    5) Hybrid gamers. How much people activatelly play on the PC and a console.

    6) Weekend gamers. How much people buys 1 title a yer. Are not new to the gamming “habit”, but don’t practice it enough, but can practice it more if some games where created that these people like.

    Withouth studies about this things, we don’t know the landscape of PC gaming.

  32. Chaz says:

    Don’t know about the hair luv, but those hot pants get the thumbs up.

  33. Grey Hunter says:

    “70 of 80 people surveyed agreed that Gillette razors are the best”*

    *Survey conducted at Gillette HQ, 50% of those questioned were female, and one of them has since been made redundant.

  34. Delusibeta says:

    Stats rule of thumb: the accuracy (assuming simple random sampling and the overall population being bigger than your sample) can be determined by 1/square root of the sample size. In the case of the 8000 figure, that’s about 0.01 either way (+-1%, in other words).

    The problem is that the sample is taken from people who signed up to give data to the NPD, which is *probably* not representative of the public (or even the gaming public) in general. For example, no non-USA people.

  35. mod the world says:

    This lady is not wearing panties!

  36. kran says:

    I used to work at Nielsen years ago. How it worked then, is we’d call people, ask them to participate, and if they did we would send them a booklet (and a certain amount of money — depending on certain criteria like age, ethnicity, geographic location) which they would fill out what shows they watched for a week, then they’d send it back.

    Those diaries would then be processed and ratings would then be projected.

Comment on this story

XHTML: Allowed code: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>