mata analysis - Can someone please help me with this mata analysis

[shivang patel]

*Data is about: The effect of Maternal Lipid Levels on Pregnant Women Without Complication in Developing Risk of Large for Gestational Age Newborn

*large-for-gestational (LGA) : Refers to a fetus or infant who is larger than expected for their age and gender.

* Import the stata data file "REVMAN.dta"

*Read Description:
* The data provides mean difference, and standard error of each samples being compared, but does not provide Effect size and standard error for the comparisons!
* An alternative is to use Cohen's d , which is the standard mean difference.
* Hint: use metan with N M Sd of LGA and nLGA!
* Type "help metan" and go to "Meta-analysis of two-group comparison of continuous outcomes, using the sample size, mean and standard deviation in the treatment and control groups"
* The metan of cohen's d can produce the effect size and standard error for you to be later used in the future analysis.
* LGA is the group affected by LGA and nLGA is the group not affected by nLGA

** Important ----> report up to 2 decimals and do not round up or down <----

ANSWER THESE QUESTIONS BASED ON THE "REVMAN.dta" file

Q1. What is the overall inverse-variance(IV)?

Q2. What is the exact name of the study with highest SMD?

Q3. Use the estimated ES and seES from the previous questions and set them as ES and SE then
perform meta regress on Weight. What is the coefficient of Weight?

Q4. Is this effect statistically significant?

[Jared Greathouse]

I smell a homework question, or even worse, a final exam question.

[Nick Cox]

The same point put differently: Please see #4 at https://www.statalist.org/forums/help#adviceextras for our policy on obvious assignments, which is a request is not to ask such questions as #1 and not to answer them.

[shivang patel]

its a homework help and I m new to this stata software. I would appreciate any help on this

[Jared Greathouse]

I understand you'd appreciate help with it, but this isn't what we do here. Let me explain why, and offer advice while we're here. Firstly, I presume you're a grad student (if not, I didn't know they gave meta-analysis courses at the undergrad level). And, if you're a grad student being taught meta-analysis, chances are that you are a quantitative researcher with at least some requisite background in stats. This means, if you plan to become a researcher, that stats is kinda your life now- it's what you do, it's your proverbial bread and butter. And as researchers, we use stats software, and learning that software is part of that process.

Because you're new to Stata, this is a good opportunity for you to learn to look through the help files of Stata commands. I'm not saying, by the way, that you should never use Statalist as a resource- Nick's code and the advice of other contributors has literally helped me from an undergrad political science student to now in my Ph.D program as a public policy student, but the only way you'll learn is by learning to interact with the software. And to learn, trust me, you must make yourself uncomfortable. I oftentimes integrate my Stata code with Python- and believe me, this makes me uncomfortable sometimes, but I still try and learn it as best I can, and the only way I can do that is by trying to work with it in the context of a real dataset.

I'm not trying to be mean in asking this, but did you actually read the directions the assignment asks for? I mean, the instructions literally say

Type "help metan"

and when you read through it, you'll find the solution you seek to questions 2-4 (I don't remember what Q1 means). My point is, I don't know you or your instructor, but assuming you have Stata 17, they're not trying to trick you on this one, they're trying to guide you to the answers without simply giving it to you.

[shivang patel]

thank you Jared Greathouse really appricaite the help.