Business Analytics Survey

[The Random NPC]

Hi Playgrounders, I've got to conduct a survey for my business analytics class and was hoping to get some volunteers. All you'll have to do is take this survey about your age and car details. Thanks in advance.

[Tvtyrant]

Objection! This doesn't offer options for not owning a car/not driving. There needs to be some options for biking, public transport, etc.

[The Random NPC]

Good point, I'll talk to my group about adding that. Thanks!
EDIT: So bad news, we can't edit the survey. For now, just use Other for non-car ownership, please.

[Manga Shoggoth]

Are you assuming that the car would be used for getting to work? I have a car, but - as I work in London - I do not use it to go to work because driving around London during rush hour is ... well, not something I do.

[The Random NPC]

We're mostly trying to see if there any commonalities across categories, such as age and distance driven. But yes, I do believe the rest of the group assumes you drive to work. I'll make sure to remind them that isn't always the case though.

[Chen]

I would seriously consider making a new survey taking these points into account. If I were a professor I’d tell you the data obtained is useless particularly since you’re posting this on an international forum. If it were limited to the US where car ownership and use is more common it may be better but you’re still going to get skewed results from this.

[snowblizz]

I would seriously consider making a new survey taking these points into account. If I were a professor I’d tell you the data obtained is useless particularly since you’re posting this on an international forum. If it were limited to the US where car ownership and use is more common it may be better but you’re still going to get skewed results from this.

Maybe that's a lesson the professor is trying to teach?

[Murk]

Eh, it's a fine survey, but if you want to use it as it is, you'll need a more rigorous selection of participants.
If you limit the survey to people who:
- Own a car
- Use that car to commute
It'll do just fine, and you might even find a nice correlation about car-using commuters and their commuting distance/gender/age/family size/income/number of cars.

I think you'll be better off keeping schmucks like me who don't own a car (or wouldn't use it for their commute anyway) out of this study. It'll be less inclusive, but stronger in the end, I think.

[JeenLeen]

I reckon part of your project can (and should) include commentary on the survey's usefulness. The points folk are raising here is a good one.

Another one: I'm in a situation where my wife and I have two cars between us, but we carpool to work. There's one car we normally use, but we do switch it up sometimes. I think an opening statement about "if you have 2+ cars, use the one you normally use for work" would help. That's the assumption I made while answering.

I also realize I don't know where my car was manufactured, only what brand it is. I reckon Hyundai and Honda have some manufacturing places in the US, but maybe it's all in Japan and shipped here. You might have some dirty data in that field from others having similar confusion or answering it based on make instead of manufacture location (assuming they aren't necessarily linked).

But I hope you take all this in good spirits. I never took the Sampling class my Statistics degree offered, but I got some summary info on it from other classes, and fine-tuning surveys is a tricky thing. I reckon learning and commenting on that is expected.

[snowblizz]

I also realize I don't know where my car was manufactured, only what brand it is. I reckon Hyundai and Honda have some manufacturing places in the US, but maybe it's all in Japan and shipped here. You might have some dirty data in that field from others having similar confusion or answering it based on make instead of manufacture location (assuming they aren't necessarily linked).

That question actually makes no sense in today's carmanufacturing industry. It only makes sense in the context of angry unionbusters in the 1980s. Or what have you.

The logistics of assembling a car and the making of the parts is a massively byzantine logistical process. Think it was a popular american truck that on average passed the border 14 times during it's assembly.
And that's ignoring that one brand can be 50%, 60%, 90%, 99% and 100% the same as another brand because of how manufacturers work together, share tehcnology or use branding in regional variations.

The correct answer is likely coing to be: "all of the above".

And it is also missing out on a number of countries. Which depending on region and brand (and over time) can change the answer completely. Yes I know there's an "Other" but how did Italy make the list and France did not e.g.?

I know we are overthinking a lot here, that's probably way beyond the purpsoe of the survey in the first place.

But I know for sure that you can score a lot of brownie points for being able to point out where your method may not be holding up to scrutiny. Nowadays for research and datagathering it's important to be aware of the limitations in the data and gathering method, becasue if not, that's where you will immediately be challenged.(*)

A logn time ago in highschool we did a project about the Bermuda triangle, we had soruces and stuff, but we failed to critically check them, it just didn't occur to us. Also there was no Internet. The teacher marked us down for that. Despite noone ever actually having taught us the rudamentaries of criticism. I always felt that was a bit unfair.

(*) I can also argue at length on the topic on not solely focusing on that aspect and essentially producing a methodically rigorous pile of nonsense, a la "water is wet".

[Murk]

[…] I also realize I don't know where my car was manufactured, only what brand it is. […]

This could be an interesting side-study as well. If you give the option "I don't know", I would expect to see a correlation between longer commutes and knowledge about the car - as in: people who regularly drive long distances take more of an interest in their car, its history and manufacturing.

But I hope you take all this in good spirits. […] Fine-tuning surveys is a tricky thing. I reckon learning and commenting on that is expected.

Also this, very much.
Pointing out the flaws in statistical research is by far the most fun part of statistical research.

[The Random NPC]

Yeah, it's a group project for a 100 level class. I'll be sure to bring up these concerns when we do the analysis later, but I think it's meant to show us how tricky surveys can be.

[The Random NPC]

The group is closing the survey on the second and I wanted to say thanks to everyone who participated, especially those that pointed out the flaws in our survey. Thank you all!

[Kiljori]

Objection! This doesn't offer options for not owning a car/not driving. There needs to be some options for biking, public transport, etc.

It's a car survey. Take your tiny carbon footprint and hit the bricks.

[Tvtyrant]

It's a car survey. Take your tiny carbon footprint and hit the bricks.

"40% of those surveyed listed their manufacturer as Other. Other is the market leader, we have to take it down."

[J-H]

I completed the survey, but the results may not reflect what you want.

My pickup truck was made in the US...but it's a Toyota.
My commute is 0-10 miles...because I work from home. I'm actually about 90-100 miles away from the office and the major metro area that I go to about 3 times a month.
My household owns two automobiles...but I also have a tractor. The tractor (used, over 25 years old) cost me more than either of our two autos did.

[snowblizz]

I completed the survey, but the results may not reflect what you want.

My pickup truck was made in the US...but it's a Toyota.
My commute is 0-10 miles...because I work from home. I'm actually about 90-100 miles away from the office and the major metro area that I go to about 3 times a month.
My household owns two automobiles...but I also have a tractor. The tractor (used, over 25 years old) cost me more than either of our two autos did.

And this shows the limitations of doing a survey. You get what you ask for, literally. Many innocent greek characters have been lost in the war against statistical unreliability,

[Knaight]

"40% of those surveyed listed their manufacturer as Other. Other is the market leader, we have to take it down."

That's before we get into "wow, Other seems to have a weirdly high proportion of small work commutes". The data is already soft because it doesn't distinguish between degrees of car ownership (personal vehicle vs. family vehicle vs. multiple vehicles), is short a N/A in a lot of places (there's a number of retired and unemployed people who don't have a commute to work/school, and that's without getting into people with multiple jobs or both work and school), no mention of leases or N/A for purchases allowing for other acquisition methods, a quick cars/household metric that is emphatically not paired with any questions about size of household...

Any data analysis this is feeding into is already going to be really messy without people who don't drive getting into the mix. Plus, even if you list as Other you run into the lack of a 0 for household cars.

[The Random NPC]

Keep it coming, the more detailed my critique, the better my grade.

[Malozing]

Good luck on your assignment.

[The Random NPC]

In case anyone was wondering, my professor didn't grade my project. But the semester is over and I got a B overall, so I don't much care!

[snowblizz]

In case anyone was wondering, my professor didn't grade my project. But the semester is over and I got a B overall, so I don't much care!

Oh come one, that doesn't fly! Make them to their job. We want to know how much we helped you.

You tell them the Forum demands to know!

[The Random NPC]

Okay, I just got my grade back. I got a 16.084 out of 20. According to the grading rubric, we had no/vague analysis and the topic was too vague.