tutorial on how to do a meta-analysis in Excel | Spreadsheet Synthesis

Подписаться 3,2 тыс.

Просмотров 23 тыс.

50% 1

Here I outline how to conduct a fixed-effect and random-effects meta-analysis on effect size data (study outcomes) using Microsoft Excel.
00:00:00 A tutorial on how to calculate the following in Microsoft Excel:
00:01:50 fixed-effect meta-analysis (weighted mean)
00:05:18 95% confidence intervals
00:09:35 homogeneity test
00:17:00 between-study variance
00:19:28 random-effects meta-analysis
#excel
#meta-analysis
#LajeunesseLab
cite this video (follow figshare link for dataset):
Lajeunesse, M.J. (2021): Fixed effect, homogeneity tests, and random-effects meta-analysis in Microsoft Excel. figshare. Online resource. DOI: 10.6084/m9.figshare.14138087
playlist of other videos using Excel for research synthesis:
• Spreadsheet Synthesis ...

Наука

Опубликовано:

14 июл 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 56

@LajeunesseLab 3 года назад

Hi all, thanks for watching! I have more tutorials using Excel for meta-analysis and systematic review practices, the full list is found here: ru-vid.com/group/PLMDcoG6cxhI0EaLoG9jG7fmf22YIUjD-o

@muhammadrana5072 Год назад

You're making research very easier SIR! Thanks

@iqrasafdar9386 Год назад

Can I get your I'd .please I am working on it and Now I am stuck

@NicoloBelavendram 2 года назад

I was looking all over the internet for meta-analysis. After your hands-on video, it became easier to see what it is all about. Excellent... and thank you very much.

@LajeunesseLab 2 года назад

Glad I could help!

@naturerelaxationmusic418 Год назад

Perfect, thanks it is more helpful!

@CalzOmon Год назад

Amazing, thank you so much

@ccahradiology7331 2 года назад

Sir, thank you for this video. This is a life saver.

@LajeunesseLab 2 года назад

Glad it helped!

@asdfghjkl7557 10 месяцев назад

This was so useful thank you so so much

@kathleenchua9946 2 года назад

Thank you so much for this

@oltishaolamo852 Год назад

That is good , as have been doing My reach seminar, Thank you dear.

@sudhabishnoi 2 года назад

Thanks for such an easy explanation 😊👍

@LajeunesseLab 2 года назад

My pleasure 😊

@Moonstonn 2 года назад

Thank you very much for this video. Whenever you said 'I'm talking through this too much", I'd respond, "no sir, you're talking just the right amount." This a godsend. I'll get some of the software you suggested but I won't forget this lesson any time soon

@LajeunesseLab 2 года назад

haha thanks, everything is recorded live and I'm always concerned that I'm taking too long -- good to know someone is listening!

@allandleeb 2 года назад

Thank you very much. This is awesome. I followed your steps and I got the same results. Could you please, add your data set for this video so people practice?

@saikatsamadder5472 6 месяцев назад

Love from India

@dom4april227A 2 года назад

Thank you for this great tutorial! Why don't use k (number of studies) in confidence interval?

@joseramos3356 Год назад

Dear, thank you very much for the video! It has been very useful for me. I am trying to perform the analysis with a dichotomous estimator and calculating the aggregate relative risk. However, the lower end of the confidence interval is negative and I have not been able to get a result that is between 0 and 1. How can I fix this? Thanks again!

@FooodConfusion 3 года назад

you are awesome

@LajeunesseLab 2 года назад

Thank you!

@aliceusa383 6 месяцев назад

Thank you! Very useful. How do we get those d and variance from literatures? For example, I got the OR=1.25 (group treatment vs control), then what is the d and variance here?

@TheF158962002 Год назад

Hi, thanks for sharing. D in the video = Cohen's d , right? How to calculate the VAR_D? thank you.

3 года назад

Again, another very well done video. What software do you use to make this presentation (the animated slides)?

@LajeunesseLab 3 года назад

I use OBS software to record live

@jonierlp 3 года назад

Thank you so much for this video! Small question, (I2 - I squared) can be calculated from Q and report like heterogenity of variance? Thank you in advance.

@LajeunesseLab 3 года назад

Yes, this Q can be used to calculate I²: I² = 100% x (Q - df)/Q.

@ogbucharles4539 2 года назад

Hi, I am struggling to run a metanalysis. It is my first time. I have continuous data with mean, sample size, and standard deviation for each study (without means for control and exposure groups.) Any ideas on the appropriate R argument suitable for this type of data?

@elenaguzman9413 2 года назад

I found your video extremely helpful. I have a better understanding on the formulas now. However, I wondering how much the formulas change if I am working with binary variables? Please let me know if I would need to do anything different. Thank you.

@LajeunesseLab 2 года назад

the formulae are applicable to all effect sizes, but things certainly get more complicated if you include predictors, see here for example: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nLOPWhO45dk.html

@reachingforthesun 2 года назад

🤣☕☕ meta analysis Olympics haha. Thx Doc

@LajeunesseLab 2 года назад

haha, I go for gold with the matrix algebra here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-nLOPWhO45dk.html

@reachingforthesun 2 года назад

@@LajeunesseLab watching👌🏽🔥 the video on meta analysis patents was neat. Stuff lay folks don't think about much. It's interesting...

@osamaabu-hammad7953 2 года назад

Thanks so much indeed. Have you uploaded these Excel work-files anywhere as templates? It would help to get hold of all those equations especially for health care workers like me who can't deal with math.

@LajeunesseLab 2 года назад

I have not uploaded any excel files as templates -- this is a good idea and I'll see what I can do

@shkhizar321 Год назад

Can you please interpret the results? It is found insignificant but what about the direction?

@user-se6yx8zr5g 2 года назад

Great tutorial! are the original effect sizes and variances (columns B and C) the original data from the papers? or what you would call "Cohen's D" and "Cohen's SD"?

@LajeunesseLab 2 года назад

Here is the data: figshare.com/articles/online_resource/Fixed_effect_homogeneity_tests_and_random-effects_meta-analysis_in_Microsoft_Excel/14138087 It's a bit complicated on what they are named, in the 1985 Hedges & Olkin book they are referred to simply as d and variance of d, which some disciplines now call Hedges' g and variance of Hedges' g. They are related to Cohen's d, but have an additional small sample bias correction factor. Hope this helps!

@user-se6yx8zr5g 2 года назад

@@LajeunesseLab Thanks for the quick response! To be more concrete- say I take data from papers that reported the outcome as means, and for each study the scale which was used was a bit different, than I should convert them to Cohen's d or Hedges' g using the means and the variance to reach the point where the tutorial starts, right? (or are columns B and C in the tutorial just the data from the paper and columns D and E the ones doing that initial standardization of the effect sizes?)

@LajeunesseLab 2 года назад

@@user-se6yx8zr5g You are correct, to get the effect sizes used in the tutorial, you would need to extract the means, variances, and sample sizes of a treatment and control group to calculate a d or g for that study. I provide examples of data extraction here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-YUcSfV0BglY.html and examples of converting these raw data into effect sizes here: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-qp2pUZZX2CU.html Hope this helps!!

@vinuvijayakumar5270 2 года назад

Hai LajeunesseLab, What about the sign (-/+) in 'd' value? Does it affect totalling the D/Var?

@LajeunesseLab 2 года назад

yes the sign is important -- it a key part of the averaging and weighting

@vinuvijayakumar5270 2 года назад

Thank You

@lakshmipriyasiva2182 2 года назад

how the variance are calculated form ODDS 95% CI

@zayarphyo5574 Год назад

I am just a beginner to do meta- analysis.Firstly, Study means citation 1,2,3 ; d means standarized mean and var D is what? I don't know how to get var D? Can you explain this?

@ChynaTropic19 3 года назад

How did you get all of those variances for your effect sizes?

@LajeunesseLab 3 года назад

When you extract the study outcomes from studies, they are often based on summary statistics like means, standard deviations, and sample sizes. With these, you can calculate effect sizes and their variances.

@ChynaTropic19 3 года назад

Yes, thank you for all of your help through the uses of the videos. I have constructed and submitted my charts already, thank you. It looked amazing. Thank you.

@ChynaTropic19 3 года назад

Also, would you happen to know another name that could be used as a title/header besides the title “conclusions” in a meta-analysis?

@ramonh.6909 3 года назад

this excel tutorial is the continuation of your meta analysis series?

@LajeunesseLab 3 года назад

Not really, in the series I mostly use R for meta-analysis, but I have colleagues who are on computers with limited software, and so I created these Excel tutorials for them since R is a headache for their systems

@Abe2957 9 месяцев назад

The papers I have are incidence/1000 population and CI of 95%, Incidence only with P-value, or simply proportion between treated and control, X2, DF, and P-value presented. There is no effect size, Stdev given. How can I get those values in order to use the excel? he presenter says i can not help you if you don't have stdev values but those values are not presented in published papers.

@anamikarana4704 2 года назад

Hello, I have seen your videos on meta-analysis. I am having certain queries. Can you please provide me your email id, so that I can have words with you. Thanks in advance.

@naturerelaxationmusic418 Год назад

Perfect, thanks it is more helpful!

@LajeunesseLab Год назад

Glad it was helpful!