Is 'Statista' a reliable source with reliable data?

[Embarrassing123]

Is there a source that has reliable data and statistics? I am in need of something reliable and 'Statista' doesn't seem reliable to mee

[BCA6010]

Using outsourced statistics is always playing with fire if you don't know exactly where and how the data was collected. Some things that have hard, provable numbers like for example the number of imported cars on a given year is probably reasonably accurate, but opinion polling and crime statistics and stuff like that is very easy to fudge. If you're doing this for a school project or work, I would definitely make sure you cite your exact source in case it gets challenged for whatever reason.

With opinions, even a barely competent pollster can word the questions or cherry pick data to get the results they want. I saw this all the time with news groups on Facebook who would re-publish polls they put out, changing the wording a bit until they got numbers closer to what they wanted. One in particular I remember was about gun control, and the original post showed that about 85% of Americans opposed it, and eventually they got it to maybe ~60%, and turned around and used their own polls that week to try and claim that some event caused 25% of Americans to change their mind about gun control.

With crime, it's harder to do with some things like bank robberies, but shootings are especially easy to fudge. Lots of published data will for example include suicides to send the number through the roof, but they'll also exclude anything considered "gang violence" when trying to push an agenda. Sometimes they'll also rope in self-defense shootings since they are often chalked up as homicides regardless of whether it was later deemed justified or not. This is especially common in cities like Chicago and Washington DC where a massively disproportionate number of firearm related murders are black on black, but they'll be swept under the rug as "gang violence" or "mutual combatants" as the DA likes to say regardless of whether it was actually gang related or not.

[Embarrassing123]

Very good point!
Yes it's for a school project. However Statista doesn't seem to reveal the source for free and i don't know if i am allowed to write Statista as my source, when they in fact aren't the source and just the website that publish the data. So i think i will do some digging and find another more reliable website that actually reveales the source for free and at the same time provides just as good statistics (which is kinda hard to find). Luckily my subject is not crime it's about gender inequality - or more specifically - gender inequality in the labour market. I'm trying to find some reliable statistics about the gender pay gap by industry and also statistics that show the labour segregation between the two genders.

[Embarrassing123]

I did find something else but it is from 2012 and this is a social science project so i don't know if it would be a bad idea to use something this old.

[WindAtMyBack]

Are you interested in some particular subject?

When dealing with statistics you should always try to check multiple sources and compare. Check the original source of the data NOT places that post data from other sources.

Check how the data was collected. Look at how they define things. Check for overlapping categories ( which often shows a bias ). If it uses sampling, check the sample, which very often is not representative of whatever group they are studying.

Check how they control for variables. This is by far the biggest problem with drawing conclusions from statistics. Just identifying all the things that could potentially affect the outcome is a major part of it, and the part that is almost always overlooked. This is particularly difficult since seemingly unrelated things can affect the outcome. This is when having some knowledge of a wide variety of topics is important.

The only way to draw conclusions is to spend lots of time digging through the data and checking multiple sources. There will almost always be discrepancies between sources. You need to identify them and resolve them.

[Embarrassing123]

Hi!
Thank you so much for this, very helpful!
The subject is social science, which is why it is important that i don't mess this up!
The topic is gender inequality in the labour market, so i found some statistics that showed the gender pay gap by occupation. I found some data that had pretty much the same numbers, should i include them as a note in my project and just write them down as second hand sources for extra credibility? And yes i was looking at this data and i found some other data that could help explain the outcome of this gender pay gap as well. However i'm having a hard time finding any overlapping categories or the sources other than the websites themselves. Some om the sides credits themselves at the sources without further explanation. I have included some data from PEW and iwpr but also from other places ): PEW and iwpr seems pretty genuine compared to stastica!

[WindAtMyBack]

Social sciences are the worst ones because it's so subjective, and it has so many variables.

One big problem with comparing pay is that there is almost no such thing as two people doing the same job. Just because they have the same job title doesn't mean they are doing the same job.

At lower level jobs, the pay is usually fixed, or has a very narrow range. Once you get above that level the most important thing for roughly eguivilent jobs, is how well the person sells themselves, and how well they negotiate. You can have two people sitting side by side with one making double what the other makes - strictly due to their ability to sell themselves and negotiate. Knowing what they are worth is part of negotiation, and is difficult to find out. As matter of fact, knowing your market value is one of the hardest things. The size of the company makes a big difference also.

There are soooo many factors. Like it's well known that shorter people make less than taller people. Women tend to be shorter, so how do they compare to men of equal height. I'm just using that for an example of the type of things you have to watch out for.

PEW claims to be non-partisan, but from what I've seen of them they aren't.

Personally I use a lot of government published data. Most, but not all government data is just straight numbers, with clear categories, and without a lot of interpreting the numbers for you. I like working with raw data like that and deciding for myself how to interpret it.

[WindAtMyBack]

Here's an example of overlapping categories.

Auto insurance companies set their rates based on risk. They use tons of statistics to determine the risk of different demographics. They know that people who live in densely populated urban areas have a higher risk of getting in an accident. They know that people in the lowest income brackets are higher risk.

However, the lowest income people usually live in urban areas. So it's an overlapping category. If the insurance company charges more for them living in an urban area, AND charges more for being low income, it would unfairly charge them too much.

Simply looking at the statistics isn't enough. You would have to also know that most low income people live in urban areas. This is knowledge about something completely different than driving. This is why having knowledge of a wide variety of topics is important. You never know what outside factor might affect things.

[Embarrassing123]

Yes it makes sense. Men also tends to have more experience and shown to be more brave which also adds to the gender pay gap. The factors i chose to look at however is statistics that show occupational segregation, family-work balance, difference in women and men working full- and part time and lastly discrimination. All these can add to the gender pay gap and at the same time reinforce each other. Sadly i do not have any room to discuss in the assignment itself whether the sources are bias or reliable or not. The teachers expect us to know whether a source is reliable or not, so i just have to judge myself and go along with it. i tried looking into some government published data but i couldn't find what i was particularly looking for.