Two-way ANOVA in GraphPad Prism

[@mc_ntg2497]

I can't believe no one has ever mentioned the narrative results checkbox to me before... Thanks for teaching me something new and useful :)

[@DoryVideo]

My pleasure!

[@DrivingABC]

Thanks a lot, James. I am from Harvard Med School, Self-learnt and used Graphpad Prism a lot. Now the labs were shut-down because of COVID-19 and your videos is a great source to learn it in a systematic way.

[@DoryVideo]

Glad you have enjoyed them! There are plenty more to come too.

[@nazaninerfany5910]

Thanks a lot, James for teaching me how to choose a parameter that best suits my analysis, helped A LOT :)

[@shaguftanaz9616]

This is quite an informative video as previously. please make a video for correction of p-value after multiple comparisons

[@robinsonvictor816]

Thanks James, I love the video.

[@DoryVideo]

Glad you love it!!

[@layth1987]

Thank you so much sir ❤❤ finally I going to do my research result ❤

[@yunsi00]

Oh man, just thank you... so well explained!

[@DoryVideo]

Glad you liked it

[@felipeduran3061]

gracias james por el video, me ayudo mucho y mas que este traducido al español, te mando muchos buenos deseos

[@CommissureforLife]

Greetings. It was really helpful for me.

[@ttp2210]

Thanks a lot. It was very helpful

[@DoryVideo]

Glad you found it useful

[@zimingli5642]

Thank you ! This really helps

[@DoryVideo]

Glad that helps

[@davitmakharadze7957]

Very useful, thank you so much!!

[@Vatierville]

Great video - so if I want to analyse my variables - looking at W.T. vs mutant strains of strep in various concentration of metal over time (absorbance) would this be better to carry out an ANOVA?

[@DoryVideo]

Sounds like a two-way ANOVA to me… two groups (WT vs Mutant) over time

[@oritsejemiyoejuetueyin3711]

Thanks for the explanation.

[@DoryVideo]

My pleasure

[@Michael-vt8ke]

For "each row represents a different timepoint", does that have to mean timepoint specifically or can it mean cell size for example? (eg each row represents the number of cells that are between 0-500μm2 , 500-1000, 1000-1500μm2 and so on)

[@AzzahAlsagoff-o6s]

I have an experiment with: - 2 different cell lines: 1.healthy cells 2. cancer cells - 2 different treatment types: 1. NK effector cells. 2. Chemo-drug (lets say this can be viewed as Drug A and Drug B for simplicity) - I analyse the viability over time. Would I also use two-way anova? If I use Grouped, would the 2 data sets be the diff cell lines? Then what are my factors: treatment type only? (there is no dose, its just treatment type). I know my Y value will be the viability. So does that mean my other independent factor is time? If I do use 2-way, which type of model should i use?

[@yelima8100]

this is so helpful, thank you!!!

[@DoryVideo]

You're so welcome!

[@Fatima-pp5ue]

ITS REALLY GOOD VIDEO, I APPRECIATED

[@DoryVideo]

Great to hear! Glad you found it helpful

[@mardisaadahmad8115]

thanks Dory

[@huizhu5302]

The most useful induction! Perfect and thank you!

[@DoryVideo]

My pleasure!

[@liselottemackens8549]

Thank you for the helpful video!

[@DoryVideo]

My pleasure!

[@julietabaglietto4418]

Hi James! I need to do an analysis of different treatments at different times but I have different individuals for each time/treatments, would this test work? Thanks!!

[@DoryVideo]

If the data are normally distributed, yes

[@mariamnaveed2681]

very useful tutorial 👍👍

[@DoryVideo]

My pleasure!

[@DrivingABC]

Hi dory, I am using version 8.4.1 which is released on March 13. 2020, but don't know why I don't have the narrative results tab after the analysis .

[@DoryVideo]

Are you SURE you are doing the analysis right

[@DrivingABC]

@@DoryVideo I have "ANOVA results" and "Multiple comparisons", but not "Narrative results". I followed your steps and just don't know why is that? BTW, I am using a Mac.

[@joaoluis97]

@@DrivingABC Hi. After choosing the ANOVA two-way, you have to click in "Options" and then in "Additional results" click to "Narrative results". I'm using the same version!

[@nickykruit]

good video heb ik veel aan

[@DoryVideo]

Glad you found it useful

[@Mkvine]

Let's say you have a control and a knockdown cell line and you stimulated them with either an "activator" or vehicle. Then within the "activator" or vehicle, you added different drug treatments to measure an effect. So you have a) Cell line b) activator and c) drug treatments. Would this be a 3-way anova?

[@DoryVideo]

Correct

[@Mkvine]

@@DoryVideo Thank you so much, that really cleared it up for me.

[@teveo5685]

Thank you!! Good video. I have GraphPad 8. I don´t have the option in "Multiple comparisons" to select: "Compare cell means regardless of rews and columns". Do I have to do two analysis? (one for rows and the other for columns?)

[@DoryVideo]

Yes that’s one way of doing it

[@brittniming-whitfield2814]

Can I use this if I have more than two groups? In my case it would be 5 instead of just wild type or mutant.

[@ahmads3799]

can I take course with you to analyse my data? please let me know

[@-maple123]

the graph pad can analyze data with Random Design complete (RAL) factorial pattern with 3 replicates?

[@SomiaBasra]

Life saving❤

[@DoryVideo]

My pleasure

[@jhonnatanocampo2688]

Hello, I have a question, when you want to evaluate the effect of a medication and take samples at two different times, shouldn't you use one-way anova for repeated measurements? I feel that the example you use in your video is similar, but you are using two-way anova, based on what did you make this decision of whether to use one-way anova for repeated samples or two-way anova? What do you recommend I use for my experiment?

[@jenniferjuarez-ayala6559]

What if your RN is 1? would that be a one-way ANOVA? because I am barely on my first trial, and I am not able to input anything less than 2 for the replicates value. Thanks!

[@inciek]

Thanks for the video. How can I set the data for the number of the same cell line from 3 patients taken at 2 different time points and 5 different doses? Thank you.

[@drkiransj8939]

Thanks James! Good video and explanation. I wonder if repeated measures two-way anova are useful for comparing growth (Absorbance) of two bacterial genotypes (with 5 biological replicates in each group) over time.

[@DoryVideo]

Yes they would be, as long as the data are parametric and pass normality test.

[@Zianaify]

What if in my control lot n=4, and protocol 1 and 2 n=6? It s not my fault, mu teacher did this

[@DoryVideo]

It’s ok, groups don’t have to be the same but if you have done a power calculation it’s worth knowing how reliable your statistics will be.

[@matthewgross2657]

Good stuff. Thanks a lot.

[@DoryVideo]

My pleasure

[@pramodinirout9771]

Thanks a lot Sir...but when p value is 0.0003 then it is correct or not...plz need suggestion

[@DoryVideo]

If you have determined before undertaking the experiment that if p

[@pramodinirout9771]

@@DoryVideo thanks

[@amandagubert9162]

Thank you so much!

[@DoryVideo]

Happy to be of help

[@JaMarPa00]

Great tutorial James. I have a question, How do you combine data from two independent experiments in Graphpad?? Lets assume I make this experiment and a week after I repeat the same experiment using the same WT cells and GPP5 cell line, how do I combine the data of the two experiments in one and analyzed statistically? Thanks.

[@DoryVideo]

Just add them to the same data sheet

[@titayaRockz1998]

When using repeated measures, when would you assume sphericity or when not to (Geisser-Greenhouse correction)? Choosing one over another gives out different p-value and its significance. I'm unsure whether or not to assume or not assume sphericity for my analysis. Would really appreciate your help :)

[@DoryVideo]

From the Prism manual: If your experimental design relies on matching rather than repeated measurements, then you can assume sphericity, as violations are essentially impossible. •If your experiment design is repeated measures, we recommend that you do not assume sphericity.

[@titayaRockz1998]

Dory Video Thank you so much! That really clear things up for me 😊

[@nehaaskari6329]

How can we make a layout if we have different concentration of drugs?

[@DoryVideo]

The layout would look different for that purpose. But would you still need a grouped format like this if all you are looking at is different concentrations, surely that would be a simpler (one way ANOVA) Analysis and, therefore, graph.

[@nehaaskari6329]

@@DoryVideo different concentrations of drugs applied at different time intervals, so we have three factors, drugs their concentrations and time

[@nogaali2274]

I used to 3species for study gene expression of 4 enzymes in two region (pollutant and un pollutant)but I don't know how used any test

[@mathildedumas2739]

Thanks a lot for this useful video ! In your example you have only 5 biological replicate for each cell line. How do you test if your data are normaly distributed before you perform the Two-Way ANOVA test for this type of experiment? Thank you for your feedback

[@DoryVideo]

You can test for normality in prism: kzbin.info/www/bejne/rXjQl3xniNqrgpY

[@iot3136]

Thank you 🙏🏽. By any chance Do you have a tutorial for 4 way ANOVA? I have 4 factors and don’t know how to enter data on GraphPad.

[@DoryVideo]

if i do one I will post it here

[@kokotampubolon26]

Thank for the video, where I can download the software?

[@DoryVideo]

www.graphpad.com they do a 30 day free trial

[@niiinjaaa3241]

Hi, thanks for the video. When it comes to anova, being two- or three-way, i don't really know how to biologically (or even statistically) interpret the "interaction" result. How would you explain it please ? To give a proper exemple, I'm working on two genes, and thus have 4 groupe (WT-WT ; WT-mut ; mut-WT ; mut-mut). In term of result, the simple mutated have an intermediate phenotype, and the double mutated a stronger phenotype. When I run the anova, I got something like "mutation1 : * * * * ; mutation 2 : * * * * ; mut1 x mut2 : ns". How to interpret that ? does it mean that the stronger phenotype in the double-mutated groupe is actually just the sum of the effect of mutation1 and mutation2, and that there is no "synergistic" effect ? To try to understand better this, i tried different set of data. If for the double mutated, i plot value equivalent to the simple mutated (thus, intermediate phenotype), THEN, i have something like "mutation1 : * * * ; mutation 2 : * * ; mut1 x mut2 : * ". The fact that mut1 and mut2 have both an effect on their own, but that they do not additionnante (as it actually does in my case) could be an indication of an epistasy between those 2 genes. Is it what the "interaction mut1 x mut2" is actually showing me ? Thanks a lot to you, who read it to the end, and thanks even more to you, who will answer me :)

[@pramodinirout9771]

Thanks a lot..I have a set of data. I want to know the significant difference in the data values using ANOVA. I have read a lot of literature similar to my work and it shows that most researcher have used different superscript letters like a, b, c along with the data. It is written " the values with different superscript letters in a column are significantly different (p

[@DoryVideo]

Glad it was helpful!

[@ferocd13]

What about the assumptions? How do you test Homoscedasticity with this program? Thanks

[@niiinjaaa3241]

Quite an old question but well, just in case : check out at 9min24sec in the video, bottom of the window there is a "diagnostic for residual" section. Select the 2 option to test Gaussian distribution and heteroscedasticity. Once you've run the program, you'll have the result of those test on the first result page of your anova (Kind of weird, because you need to run the anova to know if you are "allowed" to run the anova... but well, that's how I check)

[@ferocd13]

@@niiinjaaa3241 thank you for your answer!

[@MinhNguyen-tl3zi]

thank you very much!

[@DoryVideo]

No problem

[@elsahammerdahl8739]

Do you have any suggestions for what to do if the number of observations for each of your groups is different? I have tried to find a way to edit the number of sub columns but it seems that it automatically applies to all groups and I am not sure if there is a way to manually set them so that they can be different.

[@DoryVideo]

you cant do this but when you do the analysis pris will ignore empty cells

[@evandrovalentimdelima8484]

There is similar test to two-way Anova when the data not follow a normal distribution?

[@moumamima4407]

How did you performed the graphe sir!? Can anyone help me plz ?

[@DoryVideo]

The graph is made in prism, there are plenty of online tutorials showing this

[@zebkhan715]

Dear Sir, thanks for knowledge contribution. Kindly, your valuable suggestions required, I want to compare two groups for pre and post situation at the same time. Means 2 groups and 2 situation. any suggestion for related test or analysis. Thanks

[@DoryVideo]

You should discuss this with your supervisor.

[@zebkhan715]

@@DoryVideo Sir, I'm independent researcher. Anyhow, don't be so limited in giving suggestions. my apology

[@medicalcontent9759]

shouldnt you use repeated measure ANOVA insteade of a 2-way ANOVA

[@DoryVideo]

only if the two experiments are carried out using the SAME entities

[@nazmisoraya8599]

how homogeneity?

[@DoryVideo]

what do you mean?