Thank you ALL so much for taking the time to comment. This has been endlessly fascinating, and everyone has been so remarkably positive and enthusiastic about it! Keep reading, as there are PIE CHARTS and explanatory notes to come.
But first, pictures!
The black dress doesn't have a label - because it's the market bargain from Bangkok!
And that's the funniest thing about this whole thing: most people thought the most expensive dress - the pink one - was the cheapest.
Oh wait. I paid $109 for that Review dress. Maybe it's not that funny.
My friend Elizabeth, who is quite the most glamorous woman I know, screeched her indignation at that unfolding result via text message: "I cannot believe people think your Review dress is from Bangkok!"
She wasn't alone; a number of keen-eyed fashionistas picked the Review dress. Many pointed correctly to the photos not showing the close-up detail of the fabric and embroidery, and I do realise that seeing the dresses closer up would have made a difference.
But then, I said it was a cobbled together test. Just how cobbled together became obvious to me when I sat down to tabulate the results. I realised, using the highly technical counting mechanism of a purple Bic and the back of an old receipt, that I probably should have thought it through a bit more.
Luckily, Twitter came to my rescue, in the white-knighted form of @pappubahry, aka David Barry. David apparently does Excel spreadsheets and statistical analysis for fun, which to my mind is a far more useful skill than picking which dress is which brand. Which is good, because David got zero out of four.
David only collated the guesses that included all four dresses. Here is his summary:
- Three people picked all four - Annukya, FatCat (both from Brisbane Times) and Anonymous (from the GC blog).
- 21 people got two right, 14 got one right and 15 scored zero.
- The overall average score was 1.3, so people did better than random chance (no p-values, that's too hard for now...) Editor: What's a P-value? I didn't listen in Maths...
- Women generally picked the pink dress as the Bangkok market dress, grey as Cue, black as Coles, and by process of elimination, green as Review. However the green answers were all over the shop.
- The men generally picked pink as the Bangkok market dress, grey as Coles, black as Review and green as Cue.
Now because words about numbers generally do my head in, David kindly provided me with two pie charts, breaking down the guesses by gender. Not everyone listed their gender as part of their guess, but based on the ones that did, 36 were women, and 9 were men:
This would seem to suggest that women do have an edge over men when it comes to picking brand name frocks. Having said that, four times as many women than men took a punt, so perhaps you could just say that statistically the women had a greater chance of getting it right. However, as we've seen, very few people were able to guess all of the dresses correctly.
So what's the lesson? Well, I should've listened more in Maths, or done some research on how to compile an Excel spreadsheet. But those obvious failings aside - I hope it shows in its own little way that it probably doesn't matter what you wear, as long as you're happy with it. That may mean shelling out more, or less, or looking at other factors such as material, construction, ethical and sustainability issues, and longevity of wear.
I'm somewhat sad this experiment has come to an end. Although I do have more dresses in my wardrobe... anyone up for a round 2?
Ok! I liked the Coles dress best (probably why I thought it was Cue as I like their clothes). 0 out of 4 for me!ReplyDelete
This comment has been removed by a blog administrator.ReplyDelete
I got all 4 correct :) but I only read the article 5 mins ago, so am not included in your statistical analysis!! You've made me want to go and check out the Coles clothes though, and I'm a former brand snob (former because I finally admitted that Target has good, well priced clothes!!)ReplyDelete
The pink Review dress was the only one I got right. It's gorgeous.ReplyDelete
Sure, round 2!
Can I grab the spreadsheet off you?ReplyDelete
Have sent you an email, Sam.ReplyDelete
Upon doing some different analysisReplyDelete
log(p_i/(1-p_i)) ~ intercept + blog_i + gender_i
for p_i = total.correct_i/4 (the proportion that each person got correct)
there is no difference across BT and GC blogs and no difference between the genders (male, female, NA) when guessing all four dresses. The mean number of correct answers was 1.282 with a 95% CI of (0.784, 1.958), the expected number of correct answers is 1 (which falls within the 95% CI, so don't get too excited about our dress-picking skills).
Monte Carlo simulation (100,000) of picking from 4 dresses at random says that we expect the following proportions of people picking the following number of dresses
0: 0.37521 (3/8)
1: 0.33355 (1/3)
2: 0.24898 (1/4)
3: 0.00000 (0)
4: 0.04226 (1/24)
So we'd expect 1/24 people to get all four correct if choosing at random. 53 participants means we'd expect 2.208 to get it, and there are 3 people who get all four.
When breaking it down into individual dresses (whether they got it right or not) there was again no difference.
That is to say, respondents are no better than picking at random.
Hey David and Sam, statistics champions!ReplyDelete
I thought comments on the Brisbane Times post had closed, but they've just put through about another 60.
Do you guys want to cast your eyes over it and see if that changes anything? I'm happy to post an update later on today.
Let me know!
I think the real lesson here is statistics. That, and that your blog readers are terrible at picking out which dresses are which (conditional on your photography skills), but no worse (nor better) than those at the BT.
An interesting confounder is that being able to see others' answers might influence how you choose.
Good point because I went and looked online after reading the comments of those who had already done that.Delete
It immediately changed things for me as I could see the pink dress on the review site. But I thought the black dress was coles as they have one that looks identical on their site and I didn't see the green one there.
I am just super excited because I can buy the green one at Coles! Wheee!ReplyDelete
Wow, I totally let the team (girls) down with my 0/4. :(ReplyDelete
Can I say great buy on the green dress; I was CERTAIN it was Cue - the way the shoulders were cut. Maybe I should check out their range.
In the pink dress' defence, I picked it as the market buy because it's bright and colourful and had flowers embroidered, the kind of thing I would buy on holiday because it's cheerful, not because it looked cheap (it doesn't). holy run on sentence Batman.
Anyway, let's do another one. I have to redeem myself somehow ;)
I have to say that you are right about the Coles dresses. I went there on a shopping trip with friends to purchase food for a baby shower and walked away with 2 singlets and a dress. the dress i wore to the baby shower (boat neck with sequin detail, steel grey, A line dress with pockets). teamed up with a wide belt, everyone assumed that the dress was from CUE. it has no lining but the cut of it made it look like it was 10 times more expensive than it was. the dress was $25!! i have not yet washed it, so maybe that will be the next test?? but it definitely passed the chic test!!ReplyDelete
Good fashion as opposed to the fluff of fashion is all about the quality of the fabric and how the colours/tones of the fabric suit you. The Mix dress looks great on you Natalie as you have warm qualities, so the majority of punters did not pick it as the cheapest. The grey Cue dress looks crap on you (sorry) as it is the wrong colour for your own palette. Its about finding your own connection to fabrics and colours and textures.ReplyDelete
Sorry, but which one is the Coles dress...???ReplyDelete
I'm like Anonymous up the top I got all 4 correct! But that's no good now is it :)ReplyDelete