So Helpful - Tumblr Posts
Grimoire Prompts
Here’s a few ideas if you’re stuck on what you want to do!
Pendulum board
Crystal grid
Your favourite song, turned into a spell
Pressed flowers
Photographs of nature
Your handprint, with all the things that make you ‘you’ written inside
Pouch to hold incense
Black mirror page (Using some reflective black material)
Create your own alphabet
A letter to yourself in the future or in the past
Create an altar in your grimoire
Smells of witchcraft you love
Magical moments in your life
Dream diary
Drawings of places you have seen in meditation/ astral travel
Photos of your pets
List of reasons why you are glad to be a witch
Write a letter to a deity, or if you don’t believe in deity, to an ancestor
Colour palette tabs found from paint shops
A pie chart of significant things in your practise
Unorthodox correspondences; (examples: Socks for protection and comfort, Blue tack for creativity, doritos for enjoyment)
What would your followers leave as offerings for you if you were a deity?
Make a tarot card out of old newspaper pictures
Write a spell using the most ridiculous objects you can think off
Create a mythical creature that would be your familiar
🔥🔥🔥 WHERE WAS THIS WHEN I NEEDED IT?????????
ATEEZ Helpful guide for new Atinys
Hellooo! This is not a writing today! A new Atiny asked me to do like a helpful guide for ATEEZ! So here it is. I hope it will help many of you too! I tried to be complete as possible and linked many videos so you can get to know them a lot more! If you notice something wrong or want to add something, let me know! 😊 Thank you, stay healthy 🦋! Masterlist
REQUEST:“Okay so i’m a “new” atiny because i’m starting to stan them but i’ve listened to them before and had some really nice posts about them on my feed so I’m wondering if you can do like a little guide for ateez about their personality and which visual god is who because a girl is confused haha.”
The whole group:
Ateez / 에이티즈 (A Teenager Z, they have all teenagers wants).
8 members.
Before their official debut, the group name was KQ Fellaz.
Label : KQ Entertainment.
Debut date: 24/10/2018.
Debut MVs: Pirate King & Treasure.
Fandom: ATINY / 에이티니 (Ateez + Destiny).
US Label: RCA Records.
Lightstick: Lightiny.
Albums (April 2020): 5 korean albums (1 full album included) + 2 japanese albums.
Songs: (April 2020): 51 songs.
Fun facts about the whole group:
The group gained popularity with their Performance video I.
Korean artists are praising their energy, performance, visual, charisma on stage.
Asked their fans to vote for their title track : Wave VS Illusion.
Got their first win with Wave on 190620.
They went on their first world tour (US + Europe) few months after their debut and sold out all the stops.
They all have a position (center, rapper, dancer,…) but they all can do everything.
CRACKHEADS (this video shows it well).
Keep reading
Favorite response to a request for a mnemonic: "Roses are red, violets are blue, the long thoracic nerve innervates the serratus anterior."
~ The basic layout for the cottage.
By @theofficialsadghostclub
Literally saved my ass thanks op!
Writing Help
General Tips
The purpose of writing
Important: only you can tell your stories
Every fanfiction genre has value
HOW TO STAY MOTIVATED!
Maintaining Motivation
Reasons to Keep Writing
Stretch your hands and fingers
Analyze the stories that you like
Title creation
How to comment on your favorite fics
Moodboard sources
PIRATED FAN FICS!!!
Fanfiction is legitimate (featuring Neil Gaiman)
Angst and Horror: if you can't handle it
Publishing
Writing Smut: overcoming shame
Don't toss your work
Dialogue / Writing rules
Prologue
Paragraph breaks
Sentence length matters!
Dialogue punctuation
Scene Checklist
Body Language
Descriptions between dialogue
3 dots in a sentence (Ellipsis)
Using adverbs
Using adjectives
Coming Outs
Commonly misused words & phrases
Words to Avoid
Flirting!
World Building
World building through plot holes
Using context clues
Creating land
Government
Matriarchal society
Dystopia
Descriptions
Resources for Describing Physical things
Alcoholic drinks & cocktails
KISS SCENES
Clothing references
Sewing sources (includes history of fashion)
Hair texture
Skin color/tone
Sword / Martial arts inspo
Character Development
Character development, agency, plot
Esk*mo is a slur
Naming characters
Naming characters in other cultures
Character sheets
Character sheet
Emotional intelligence
Human body limits
Making threats and the third option
Bartenders
Children
Kids
Asian characters
Muslim characters
Slavic characters
Russian names
Mixed characters
Bilingual characters
Blind or Visually Impaired
Morally grey
Dangerous female characters
Liars
Romance development
Healthy Relationship
Sacrifices
Redemption arc
Redemption arc 2
Regret / Remorse
Prompts/one-liners
Prompt masterlist
Another prompt masterlist
A third prompt masterlist
Best friends to lovers
Mob Boss
Prophecy of lost child
Super power of truth
Enemies to lovers
Enemies to lovers: Reasons to hate
Enemies to lovers: Getting together
Enemies to lovers Prompts!
Enemies to lovers: Fake dating
Past-life Enemies to lovers
Late-night wandering
Compliments
Touch-starved
Romance/physical intimacy prompts
Romance: little acts of love
Smut
Reputation
Break-ups
More sources
Masterlist: body language, words, translator
Masterlist: prompts, LGBTQ+, NSFW Advice
Meme: fanfiction problems
Mr. Darcy (Pride and Prejudice)
Bechdel Test
Part 2 ->
Thank you so much sun :((((
You're so sweet and this is really helpful❤❤ I'll definitely check them out! Really, thanks for taking the time to share these with me
Also pls I've been seeing you struggling with the heat😂😂 keeping you in my thoughts & prayers fr haha
Lollapalooza was so wild😭😭 ARMS EVERYWHERE, ALL AT ONCE😭
Hi hi!!! I hope you're well ❤
Pls may you suggest any good apps/websites for giffing because all the ones I know have such terrible quality😭
Also how are you holding up after skz at lolla😂✋
hii im doing okay (surviving hot summer) thank you I hope you're well too<3
before I started using ps , I did do some research like that was like a while ago and I remember coming across PHOTOPEA , I didn't use it but there is a great tutorial here :)
this blog also has some good lists :)
and this website has some new free alternatives to ps that work pretty similar :)
also this is me "holding up" after skz lollapalooza:
I'M SO IN LOVE OMG
REVERSE TROPE WRITING PROMPTS
Too many beds
Accidentally kidnapping a mafia boss
Really nice guy who hates only you
Academic rivals except it’s two teachers who compete to have the best class
Divorce of convenience
Too much communication
True hate’s kiss (only kissing your enemy can break a curse)
Dating your enemy’s sibling
Lovers to enemies
Hate at first sight
Love triangle where the two love interests get together instead
Fake amnesia
Soulmates who are fated to kill each other
Strangers to enemies
Instead of fake dating, everyone is convinced that you aren’t actually dating
Too hot to cuddle
Love interest CEO is a himbo/bimbo who runs their company into the ground
Nursing home au
Not only a follower appreciation giveaway, but also a signal boost for the existence of @findingwinteriron, who put in huge efforts to help people find the stories they loved but lost! One huge thank you to all of the work they do: we love you for it!
The First Giveaway!
Hello, FindingWinteriron friends! We recently crossed the 500 follower mark, so I thought it’d be a good excuse for me to give things away time for us to have our first follower appreciation giveaway! Let’s start with the goods:
The Prizes!
Option 1: A 3,000ish word fic from @aurumacadicus (aka Reioka)! If you have a prompt for the incomparable AurumAcadicus, this is your chance! They have been on hiatus for a few months, so this is really an excellent opportunity! Examples of their work can be found on their tumblr or on AO3: https://archiveofourown.org/users/Reioka/
Option 2: A Happy Couple Tony/Bucky Feltie pair by @tisfan! These felties are hand crafted, but should look something like this when complete:
Look at how cute they are! Bask in their adorableness! For more examples of @tisfans’s work, or to buy your very own because you just CANNOT WAIT to own your own feltie Avengers, hit up their etsy shop here: https://www.etsy.com/shop/FeltAvengers
Option 3: A cross stitch by @rebelmeg! You can choose any combination of Bucky and Tony from the below image, and rebelmeg will stitch them up for you!
Will you choose Iron Man and the Winter Solider? Maybe Bucky in the Red Henley of Sexiness and Tony in his workshop clothes? It’s all up to you!
And now, you ask, how can you win such bounty?
The Rules
First, you must be following @findingwinteriron. This is, after all, a follower appreciation giveaway.
Giveaway blogs are not eligible (yes, we will check)
You can earn up to three entries:
One entry for linking this post;
One entry for reblogging this post; and
One entry for adding your thoughts AND tags about what @findingwinteriron does (things like “I want to win” aren’t eligible. Tags like “fic finder” and “bucky/tony” are fabulous)
Winners will be chosen at random. The first winner we choose will be offered a choice the three prizes, the second will choose fro the remaining two, and the third will be offered the remaining prize.
You must be willing to supply a name and mailing address if you choose the felties or the cross stitch.
Creators supplying prizes as well as our friends outside the US are eligible to win.
Be polite and patient. These things sometimes take time.
Ignore my typos. Seriously. I’m physically incapable of typing more than three words in a row without messing one of them up.
So that’s it! We’ll keep entries open until noon central time on Sunday, July 7 and start contacting winners soon after.
If you have any questions, please send us an ask off anon, and we will reply privately.
Thanks and good luck!
- S.
Jared and Jensen in Conventions throughout the years ♥
A while ago, I posted a thread on twitter to help myself and other people to find easily the convention we were looking for, using the boys' outfits to identify where the convention had happened. This thread has been really useful to me and now, with the uncertainty of twitter's future, I'm reposting it here.
I also have another blog on tumblr where I post all the convention media I find: @j2panelsmedia
I hope this will help some of you as it did to me :)
I'll be eventually editing this post to add the most recent conventions.
P.S.: I only added the conventions both Jared and Jensen participated. If you find any mistakes, please let me know!
Always looking forward to new content of them together ♥
P.s.: It has come to my attention that I forgot to add the 2010 JIBCON, so you won't find it here.
My mom: are you ok?
Me: *shrugs*
My emotionally abusive household member: are you depressed?
Me: yeah…
My emotionally abusive household member: life just sucks, there’s not much you can do about it...
Let's talk about sex theft & plagiarism...
Hello lovelies,
Seeing as there's a current spate of theft & plagiarism going around the fandom (and sadly, probably will again in the future, no doubt), I've compiled some research and some info on what you can do about it, whether you've been stolen from yourself, or just want to support the wider community that have been/are being targeted.
☝️This is written referring to Pedro Pascal character works, as this is the fandom I'm in, but it applies to any and all fandoms where fan fiction in any form of media is created.
And to anyone reading this who has knowingly and deliberately stolen work - be that fiction or art - and is planning to either do so again, or is refusing to remove works out of spite or whatever - just know that we are a community here, and we look after one another.
You are, and always will be, vastly out of your depth if you think this behaviour is acceptable. Do better.
Help! My work has been stolen! What can I do?
Okay, first of all take a minute. You're angry, insulted and probably damn right fuming! As you should be. This is YOUR work, after all. Take a breath or two and let's see what we can do.
I know, the gall of some people, right? It's getting tiring...
First things first. Any written word that comes from YOUR brain and hand, be that on paper in a note book, or that you type into a Word/Google doc, belongs to YOU. Your old journal from when you were 14? Yep, that's protected.
But, my work is in the Public Domain... isn't it fair game because it's fan fiction?
No, it's not in the Public Domain, and no, it does not matter if it's fan fiction or another form of fiction.
So Jett, what is the Public Domain?
Okay, let's get wordy and speak legal here, with a little help from my trusty dude Google...
"The Public Domain is a collection of creative works that are not protected by intellectual property laws, such a copyright, patents or trademarks. This means that no one owns these works, and anyone can use them without permission or citing the original author. Public domain works can be used, adapted, reproduced, or distributed for commercial or noncommercial purposes.
"A work is generally considered to be within the public domain if it's ineligible for copyright protection or its copyright has expired. Public domain works can serve as the foundation for new creative works and can be quoted extensively.
"Works created before the existence of copyright and patent laws also form part of the public domain. For example, the Bible and the inventions of Archimedes are in the public domain. When trying to determine whether an unpublished work is in the public domain, find the date of the death of the author. This can be found searching the free Social Security Death Index. An unpublished work created before 1978 enters the public domain 70 years after the author's death or in 2002, whichever is later..."
Blah, blah, blah. You get it.
But wait, you just said that my work is automatically copyrighted to me? I'm confused! Is my work in the Public Domain or not as I post it on a publically accessed site...? 🤯
Fanfiction is NOT in the Public Domain. That's completely different to something being on a publically accessed website.
Fanfiction authors own the copyright to their original contributions to a fan work, but not to the underlying work it's based on.
For example, you wrote a version of Joel Miller as your dad's best friend...
(he loves the attention really)
... but you do not own the character Joel Miller legally. However, if the source work has entered the Public Domain, it's legal to make money from a version of it. (The Last Of Us or the character Joel Miller is NOT in the Public Domain, FYI. So no payday for you.)
So, my work is protected by copyright? But I haven't registered it anywhere officially!
Yes, and you don't have to, officially. As I mentioned, the moment you write it/type it out, it's copyrighted to you.
So, what exactly is a copyright? 🤔
Fan fiction is generally regarded as derivative work, a piece based on original work but which has originality of its own (Joel Miller as your dad's best friend again...) and is not merely a copy of the work on which it is based. If derivative works are created lawfully, they're also subject to copyright in their own right, which is what most fan fiction is.
In simple terms, a derivative work is a work that is based on (derived from) another work; for example a painting based on a photograph, a collage, a musical work based on an existing piece or samples, a screenplay based on a book. A fan fiction of The Last Of Us - all examples of derivative work.
So, fanfiction is a derivative work and NOT in the Public Domain?
Yes! You are creating your own work that is not a direct copy, but an influence from the original work. If you copy the original work, word for word, you would be in breach of copyright and plagiarism.
Ah yes, plagiarism... So, what does that mean as it's been thrown around a lot?
Plagiarism is the practice of taking someone else's work or ideas and passing them off as one's own.
So, if I took your posted Joel Miller story as your dad's best friend and posted it word for word, I would be plagiarising your fan fiction.
If I wrote Joel Miller as your dad's best friend too, but had no direct copy of your exact words or idea in it's entirety, then I am not plagiarising.
Remember that essay your teacher had you write that you copied and pasted from Google, or the homework you copied from a classroom chum and submitted it for your grade? Yep, that's plagiarism, buddy.
Basically, when you take someone's work and pass it off as your own work, that you say you created, it's plagiarism.
But, there are lots of fan fictions about Joel being your "dad's best friend", for example...
I mean, this is just an example, but anyone can have the same idea as you.
Ideas are not regulated or able to be legally challenged, unless they are officially registered as patented. You can't really patent a fan fic as the work it's based from already exists.
Anyone can write Joel as your dad's best friend. But if they write the story exactly word for word, then it's direct plagiarism. And that's a big no-no.
Is re-blogging someone's work plagiarism?
No. Tumblr gives you the option to share the original post by re-blogging and it's highly encouraged as a way to share the things you love and to support writers/artists by getting their work out there for a wider audience to enjoy.
However, if you cut and paste it onto your own post, then it's plagiarism.
Is plagiarism illegal then?
Yes, and no. Depends. Back to good ol' Google...
"Plagiarism is not illegal in the United States or Europe in most cases, but it can be considered a violation of ethics or honor codes. Depending on the circumstances, plagiarism can result in disciplinary action from a school or workplace, or even criminal charges.
BUT, and that is a big but that does not lie, "plagiarism CAN be illegal and may result in legal action if it violates an author's intellectual property rights, such as copyright, patent, or trademark."
Ding! Ding! Ding! Fan Fiction! FAN FICTION IS COPYRIGHTED TO YOU, THE CREATOR!🚨 Therefore, plagiarism is in breach of that.
"For example, a copyright owner can sue a plagiarizer in federal court for copyright infringement. The plagiarist may be required to pay the copyright owner for any losses, as well as attorney's fees. In some cases, plagiarism can also be considered a felony, which may result in a prison sentence." (Yes, it has happened!)
But how can I track down these people online to sue them, that would be impossible, right?
Well, yes, to some extent. You would probably need a lot of time and astute experience in successfully - and legally - tracking people down online. However, anonymous people online HAVE been successfully found and sued for stealing other's work and passing it off as their own in the past. Yes, even for fan fiction.
All it takes is for them to piss off the wrong person, right?
You'll need to make sure you have solid proof that you are the copyright holder and the work originated with you. You have the right to take legal action and launch a copyright lawsuit and sue for damages in a small claims court if you so wish.
So, how do I obtain proof? 🕵️♀️
In this day and age everyone has a device in their hands to provide evidence - it's your phone/laptop.
Do you know why online/digital docs have dates stored on the properties tab/settings? That's so the creation date can be logged. Same as when you take a photograph, the photo automatically has a date and time stamp on it - this is to record when exactly that piece of media was created.
And we all know, that media never fully disappears...
☝️The only person that can claim legal ownership and copyright to any media, whether posted online or not, is the original content creator of that media.
There are some other ways you can protect and store your work for evidence too:
Register your work. You can register your copyright with your country's Copyright Service, usually online or by post. Registration can provide a verifiable record of your work, deter infringement, and speed up dispute resolution. The filing fee for copyright registration ranges anywhere between $35 to $55 approx. per work submitted.
Mark your work. You can identify yourself as the copyright owner by marking your work with the international copyright symbol ©, including your name, and including the year of creation. You can also include a correctly worded notice that states the work is protected under law somewhere eon your blog/website. (You can find example templates online you can use.) If you're an artist, your signature or a digital watermark can be added to your work, along with the date you created it.
Keep records. You can keep records of when and how your work was created, what materials were used, and where and when it was published. This could include publication records in newspapers, magazines, websites, social media, or other platforms. You can update and save Word and Google docs. You can also keep early drafts, synopses, rough recordings, sketches, and deliberate mistakes. Back your stuff up to a cloud based drive or back up server/hard drive.
Store copies. You can store copies of your work in a way that provides a time stamp, such as by emailing or saving it electronically. You can also leave a dated copy of your work with a bank or solicitor, or post a copy of your work to yourself by special delivery and leave the package unopened. This is commonly known as a 'poor man's copyright.'
Okay, so even though I've protected my work, and we've established it legally belongs to me under a copyright, and I can prove it's mine, what can I do about someone posting it on Tumblr or somewhere else and claiming it's theirs?
Almost 100% of the time, the person plagiarising you is doing it for shits and giggles. They aren't intelligent enough to understand how it works, affect's the content creator, and what they're doing is morally, ethically and legally wrong. And they don't care. Their immaturity painfully shows.
Also, dates don't lie, bub. And hips. Their copy will almost, always be dated AFTER your original post.
So, anyone claiming it's theirs is frankly declaring themselves publically as a simpleton for all to see.
☝️First thing's first, only the person who owns the copyright to the work can file a report for plagiarism/copyright breach. Tumblr offers this option in the report a user setting.
You can reach out to the person advising them politely you are the original content creator and hold the copyright to work and ask them to remove it. Sometimes they do, and sometimes they don't.
In instances where they refuse, it's important to disengage from further communication at that point and go through the reporting option with Tumblr, or any website's reporting tool that your work has been posted to.
Block that person and don't enter into any discourse with them.
This will only just encourage them further and is exactly what they want. They want the attention and to be revered. Don't give it to them.
They may have never engaged with you previously and you won't always know who they are, so try not to take it personally. The issue is theirs and it's often a jealously issue or because, frankly, they're bored and have nothing else better to do than wind people up online. I know, it's sad...
You are better than that. Don't stoop to their level.
I've reported them, but my work is still up.
Keep reporting them. It may take a little while for Tumblr, or the site elsewhere, to act and delete their account. You can have your friends and supporters also report them too for harassment if they refuse to remove it.
Plagiarism is actually a form of harassment, FYI. It's the deliberate act of knowingly taking something from someone else and passing it off as your own to cause deliberate, malicious intent.
Basically, don't feed the fire oxygen. 🔥
If you don't give them any attention, they'll likely get bored of sending themselves ASK's in support of their silly actions and disappear. You can notify your followers and supporters that your work has been stolen and not to engage with the stolen work on any other blog/website, except for yours.
Most, if not all, of your followers will recognise your work, they love it after all, so they're your best allies against this.
However DO NOT gang up on, send threatening messages or Anons to the person plagiarising. This will only give them motivation to report YOU for harassment to them. They will have to prove this, of course.
Ultimately, unless you're prepared to take the person to court, providing you have the means and time to do so, then it becomes a waiting game. 🕑
I've left my blog up but don't contribute/post to it anymore/I've deleted my blog, but my work still exists in the form of re-blogs around the site - what can I do?
Whether you still actively post or have left a fandom/blog to go "dark", your work is NOT fair game for anyone to steal/use/pass off as their own, without your express written consent.
Your work is still protected by your owner's copyright, even if your blog is deactivated, or you posted that story over ten years ago. It's still YOUR work.
And as long as it exists in any from, it still rightfully and legally belongs to you. Until it becomes part of the Public Domain, which will happen approx. 70 years after your death... by then you would've hoped they've given up... lol. 🙃
Tumblr actually puts the date next to your username when you deactivate, so anyone with half a brain will see that date and can compare it to the date of your original blog post. Like I say, dates never lie. 😎
And, if you were to successfully take someone to court over plagiarism/copyright theft, their rationale of "well, their blog is fair game because they no longer post on it, so what's the problem? Yap, yap, yap..." won't stand up as a valid defence.
I know, it's laughable... these are the children of our future, folks. Sigh...
How can I protect myself in the future then?
You can make it clear on your blog/website that your work is officially protected by copyright and belongs to you. Register your work if you can. Don't engage with the thieves, report them and block them. Spread the word in the fandom so people can be on alert. And then carry on living your best life.
REPORT & BLOCK! REPORT & BLOCK! REPORT & BLOCK!
I hope this has been useful/helpful if you made it this far. Please consider re-blogging for a wider reach. Solidarity, peeps. ✊
Keep being a King, stay kind and keep it creamy. These people are not worth your enjoyment and talent being diminished.🖤
Everyone asked for it, so here it is :)
hope it helps!
SOME BG3 CHARACTER ART THINGS that I struggled to capture/notice early on that I now pass along:
1. Gale's nose is bigger than you think it is, but it's not pointy or sharp. He has a light complexion, but not a pale one. His eyes are big, but are both down turned and very hooded.
2. Wyll is scarred, not wrinkled. He is 24. His eyes are bigger than you think, and fold a bit at the inner corner. His cupids bow is Perfect and defined.
3. Astarion doesn't actually have much of a cupids bow at all, but the philtrum is very defined. His eyes come to very feline points on the inside. His shoulders are broader than you think.
4. Karlach ALSO has a picture-perfect cupids bow, and a very soft brow ridge (the slope where her nose and brow bone continue to her forehead is shallow)
5. Shadowheart looks exactly like you remember her to look like. Do not second guess yourself.
6. Lae'zel has a VERY pouty mouth. And big wet puppy eyes. Do not let her fool you. She has the facial ratios of Betty Boop.
7. Halsin, while of course very rugged, still has the "elf" ratios to his face. He has dainty and sharp cheekbones, and the same "feline" point in the corners of his eyes, same as Astarion
8. So does Minthara. Both Minthara and Halsin have very thin lips.
posted some Astarion tutorial on my patreon. I made it on twitter long long ago, so probably there are people who never saw it ♥️
Bonus post: Stats 101 - testing data for normality & significance tests for categorical and continuous variables.
Understanding and analysing data can be a tremendously daunting task, so I thought I would put together a simple go-to guide on how to approach your data, whether it be numerical or categorical. 📈📊
This post will cover:
Types of data
Contingency tables and significance tests for categorical data
Testing for normality in continuous data
Significance tests for continuous variables
NB: Remember to keep your data organised, especially if you are using software packages like ‘R’, MATLAB, etc.
Before I move on, I would like to thank the University of Sheffield core bioinformatics group for most of the content below. 💡
Types of data
There are two main types:
Numerical - data that is measurable, such as time, height, weight, amount, and so on. You can identify numerical data by seeing if you can average or order the data in either ascending or descending order.
Continuous numerical data has an infinite number of possible values, which can be represented as whole numbers or fractions e.g. temperature, age.
Discrete numerical data is based on counts. Only a finite number of values is possible, and the values cannot be subdivided e.g. number of red blood cells in a sample, number of flowers in a field.
Categorical - represents types of data that may be divided into groups e.g. race, sex, age group, educational level.
Nominal categorical data is used to label variables without providing any quantitative value e.g smoker or non smoker.
Ordinal categorical data has variables that exist in naturally occurring ordered categories and the distances between the categories is not known e.g. heat level of a chilli pepper, movie ratings, anything involving a Likert scale.
Contingency tables & significance tests for categorical variables
Contingency tables (also called crosstabs or two-way tables) are used in statistics to summarise the relationship between several categorical variables.
An example of a contingency table:
A great way to visualise categorical data is to use a bar plot/chart, which looks something like this:
There are two main hypothesis tests for categorical data:
Chi-squared test
Fisher exact test
Chi-squared test:
Compares the distribution of two categorical variables in a contingency table to see if they are related e.g. smoking and prevalence of lung cancer.
Measures difference between what is actually observed in the data and what would be expected if there was truly no relationship between the variables.
Fisher exact test:
Is used instead of Chi-squared when >20% of cells have expected values of <5, or any cell has a count of <1.
If you want to compare several contingency tables for repeated tests of independence i.e. when you have data that you’ve repeated at different times or locations, you can use the Cochran-Mantel-Haenszel test.
More detail:
In this situation, there are three nominal categorical variables: the two variables of the contingency test of independence, and the third nominal variable that identifies the repeats (such as different times, different locations, or different studies). For example, you conduct an experiment in winter to see whether legwarmers reduce arthritis. With just one set of people, you’d have two nominal variables (legwarmers vs. control, reduced pain vs. same level of pain), each with two values. If you repeated the same experiment in spring, with a new group, and then again in summer, you would have an added variable: different seasons and groups. You could just add the data together and do a Fisher’s exact test, but it would be better to keep each of the three experiments separate. Maybe legwarmers work in the winter but not in the summer, or maybe your first set of volunteers had worse arthritis than your second and third sets etc. In addition, combining different studies together can show a “significant” difference in proportions when there isn’t one, or even show the opposite of a true difference. This is known as Simpson’s paradox. To avoid this, it’s better to use the Cochran-Mantel-Haenszel for this type of data.
Testing for normality in continuous data
The first thing you should do before you do ANYTHING else with your continuous data, is determine whether it is or isn’t normally distributed, this will in turn help you choose the correct significance test to analyse your data.
A normal (also known as parametric) distribution is a symmetric distribution where most of the observations cluster around the central peak and the probabilities for values further away from the mean taper off equally in both directions. If plotted, this will look like a symmetrical bell-shaped graph:
A standard deviation (SD) can be calculated to measure the amount of variation or dispersion of a set of values from the mean. The main and most important purpose of this is to understand how spread out a data set is; a high SD implies that, on average, data points are all pretty far from the average. The opposite is true for a low SD means most points are very close to the average. Generally, smaller variability is better because it represents more precise measurements and yields more accurate analyses..
In a normal distribution, SD will look something like this:
In a normal distribution, skewness (measure of assymetry) and kurtosis (the sharpness of the peak) should be equal to or close to 0, otherwise it becomes a variable distribution.
Testing for normality
Various graphical methods are available to assess the normality of a distribution. The main ones are:
A histogram, which will look something like this:
Histograms help visually identify whether the data is normally distributed based on the aforementioned skewness and kurtosis.
A Q-Q plot:
Q-Q plots allow to compare the quantiles of a data set against a theoretical normal distribution. If the majority of points lie on the diagonal line then the data are approximately normal.
and…
A box plot:
A box plot is an excellent way of displaying continuous data when you are interested in the spread of your data. The thick horizontal bar indicates the median, the top and bottom of the box indicate the interquartile range, and the whiskers represent the spread of data outside of this interquartile range. The dots beyond the whiskers represent outliers, which represent observations that are distant from other observations.
A disadvantage of the box plot is that you don’t see the exact data points. However, box plots are very useful in large datasets where plotting all of the data may give an unclear picture of the shape of your data.
A violin plot is sometimes used in conjunction with the box plot to show density information.
Keep in mind that for real-life data, the results are unlikely to give a perfect plot, so some degree of judgement and prior experience with the data type are required.
Significance tests
Aside from graphical methods, there are also significance tests, which are used to test for normality. These tests compare data to a normal distribution, whereby if the result is significant the distribution is NOT normal.
The three most common tests are:
Shapiro-Wilk Test (sample size <5000)
Anderson-Darling Test (sample size > or = 20)
Kolmogorov-Smirnov Test (sample size > or = 1000)
Significance tests for continuous variables
A quick guide for choosing the appropriate test for your data set:
t-test - normally distributed (parametric) data
There are three types of t-test
One sample t-test: Compares the mean of the sample with a pre-specified value (population mean) e.g. if the average score of medical students in UK universities is 72 and you want to test whether the average score of medical students in your university is higher/lower, you would need to specify the population mean, in this case 72, when running your t-test.
A two-sample t-test: Should be used if you want to compare the measurements of two populations. There are two types of the two-sample t-test: paired (dependent) and independent (unpaired). To make the correct choice, you need to understand your underlying data.
Dependent samples t-test (paired): Compares the mean between two dependent groups e.g. comparing the average score of medical students at the University of Sheffield before and after attending a revision course, or comparing the mean blood pressure of patients before and after treatment. Independent samples t-test (unpaired): Compares the mean between two independent groups e.g. average score of medical students between University of Sheffield and the University of Leeds, or comparing the mean response of two groups of patients to treatment vs. control in a clinical trial.
There are several assumptions for the independent (unpaired) t-test:
The t-test assumes that the data has equal variance and relies on the data to be normally-distributed. If there isn’t sufficient confidence in this assumption, there are different statistical tests that can be applied. Rather than calculating and comparing the means and variances of different groups they are rank-based methods. However, they still come with a set of assumptions and involve the generation of test statistics and p-values.
Welch t-test, for instance, assumes differences in variance.
Wilcoxon test (also commonly known as the Mann-Whitney U test) can be used when the data is not normally distributed. This test should not be confused with the Wilcoxon signed rank test (which is used for paired tests).
The assumptions of the Wilcoxon/Mann-Whitney U test are as follows:
The dependent variable is ordinal or continuous.
The data consist of a randomly selected sample of independent observations from two independent groups.
The dependent variables for the two independent groups share a similar shape.
Summary of the above:
ANOVA - normally distributed (parametric) data
Like the t-test, there are several types of ANOVA tests:
One-way ANOVA:
Equivalent to the independent t-test but for > 2 groups. If you want to compare more than two groups, a one-way ANOVA can be used to simultaneously compare all groups, rather than carrying out several individual two-sample t-tests e.g. to compare the mean of average scores of medical students between the University of Sheffield, the University of Leeds, and the University of Manchester.
The main advantage of doing this is that it reduces the number of tests being carried out, meaning that the type I error rate is also reduced.
Two-way ANOVA: 2 categorical (grouping variables) e.g. comparing the average score of medical students between the University of Sheffield, the University of Leeds, and the University of Manchester AND between males and females.
Repeated measures ANOVA
Equivalent to a paired t-test but for >2 repeated measures e.g. comparing the average score of medical students at University of Sheffield for mid-terms, terms, and finals.
If any of the above ANOVA tests produce a significant result, you also need to carry out a Post-Hoc test.
Post-Hoc test e.g. Tukey HSD
A significant ANOVA result it tells us that there is at least on difference in the groups. However, it does not tell us which group is different. For this, we can apply a post-hoc test such as the Tukey HSD (honest significant difference) test, which is a statistical tool used to determine which sets of data produced a statistically significant result…
For example, for the average scores of medical students between the University of Sheffield, the University of Leeds, and the University of Manchester, the Tukey HSD output may look something like this:
This shows a significant difference between medical students in Manchester and Sheffield and between Leeds and Manchester but not Leeds and Sheffield.
Kruskal Wallis and Friedman tests
Data that does not meet the assumptions of ANOVA (e.g. normality) can be tested using a non-parametric alternative. The Kruskal-Wallis test is derived from the one-way ANOVA, but uses ranks rather than actual observations. It is also the extension of the Mann-Whitney U test to greater than two groups. Like the one-way ANOVA, this will only tell us that at least one group is different and not specifically which group(s). The Post-Hoc Dunn test is recommended, which also performs a multiple testing correction. For the Friedman test, you can use the Wilcoxon signed-ranks Post-Hoc test. And that is your go-to guide to on how to approach your data! I really hope you find it useful; it definitely helps clarify things for me. ✨
GOOD LUCK!
RadenWA is honestly a hero for these
they're got even more than these, too!