2nd Sep, 2021
Catholic University of Applied Sciences Munich, Munich, Bavaria, Germany
Question
Asked 20th May, 2016
Do effect size heuristics for standardised beta's exist?
Are there any justifiable method/heuristic for assessing the rough effect size (e.g. small, medium or large) of standardized beta coefficients from multiple regressions and path analysis?
Most recent answer
Christian Young Christian, since I have a similar issue, which "rules of thumb" did you use in the end for the standardised regression coefficients?
I took a look at your paper but I could not find anything for the Betas. It says "We followed conventional rules of thumb for effect sizes [29] and deemed medium effect sizes as: Cohen’s d = .5, zero-order correlation coefficient r = |.3|, and odds ratios = 2 or .5; large effect sizes were defined as Cohen’s d = .8, zero-order correlation coefficient r = |.5|, and odds ratios = 5 or.2. All other statistics were interpreted within the context of the study."
Elissa J Hamlat Elissa, did you find a workable solution?
Popular Answers (1)
In that case you may want to look at
Standardised regression coefficient as an effect size index in summarising
findings in epidemiological studies; Epidemiology Biostatistics and Public Health - 2013, Volume 10, Number 4
6 Recommendations
All Answers (10)
Dear Christian,
"Cohen’s d is a good example of a standardized effect size measurement. It’s equivalent in many ways to a standardized regression coefficient (labeled beta in some software). Both are standardized measures-they divide the size of the effect by the relevant standard deviations. So instead of being in terms of the original units of X and Y, both Cohen’s d and standardized regression coefficients are in terms of standard deviations."
2 Recommendations
You may want to have a look at this current debate on how to compare effect size in regression in favour of unstandardized coefficients
1 Recommendation
Thanks for your answers. To clarify, I am conducting a systematic review and am hoping to give a rough approximation of effect size based on multiple types of statistics (r’s, odds ratios, beta’s, b’s, etc.) I have conventions for correlations, t and F tests, and odds ratios. But am struggling to find and effect size convention (small, medium, large) for standardised regression coefficients (beta) as reported using multiple regressions and path analysis. As I understand, beta is the standard deviation change in the DV with one standard deviation change in the change in the IV, holding all other IV’s constant. Is there a justifiable method for assessing the effect size of beta in this context? Thanks.
2 Recommendations
In that case you may want to look at
Standardised regression coefficient as an effect size index in summarising
findings in epidemiological studies; Epidemiology Biostatistics and Public Health - 2013, Volume 10, Number 4
6 Recommendations
Christian Young Did you ever find a satisfactory solution? I am conducting a meta-analysis and have a similar issue. Thanks.
Christian Young Elissa J Hamlat I've used Acock (2014) (see this post for details: https://stats.stackexchange.com/questions/235110/how-to-evaluate-effect-size-from-a-regression-output/235259)
"Acock (2014) also argues that they can be interpreted similar to correlations: β^∗<0.2β^∗<0.2 is considered a weak, 0.2<β^∗<0.50.2<β^∗<0.5 moderate, and β^∗>0.5β^∗>0.5 strong effect (p.272)"
Daniel P. Moriarity Thanks Daniel! I've also found an article that suggests a formula for conversion from β to r. However, it's not helpful in the case of large β, as some β can be over 1.
My goal is to ultimately convert all to Hedges g' for meta-analysis.
Elissa J Hamlat I never really did, though the paper Daniel P. Moriarity has linked to above looks useful and is very similar to rules of thumb that I used in the end.
Similar questions and discussions
Scientists Support Ukraine
Like so many, I am shocked and saddened at seeing war break out in Europe. My thoughts – and those of the ResearchGate team – are with the people of Ukraine and everyone affected.
ResearchGate is an international company, whose purpose is to enable scientists across the world to work together openly and collaboratively, regardless of borders or nationality. We have people from over 40 countries on our staff of around 200, and being based in Berlin, we are profoundly aware of the human cost of conflicts, the echoes of which have shaped and scarred our home city. We join with the international community in condemning the actions of the Russian state.
We have been asking ourselves: What can we do?
From today, we will offer free advertising space worth $2.5 million on our network to humanitarian organizations working to respond to the crisis. ResearchGate benefits from over 50 million visitors every month, and we hope this initiative can help raise funds and awareness for those organizations that are having direct impact and need support.
We also want to use our platform to highlight the response from the scientific community. Personally, I have found the messages of support from scientists everywhere to be truly heartfelt, and I would like to highlight some of the community initiatives I’ve seen here:
- Science for Ukraine provides an overview of labs offering a place for researchers and students who are affected to work from, as well as offers of employment, funding, and accommodation: https://scienceforukraine.eu/
- Labs supporting Ukrainian Scientists is an expansive list of labs and PIs offering support at this time. Submissions go here: https://docs.google.com/forms/d/e/1FAIpQLSeRGe5Da_b6GGyC6VT7CLGViGs06SzeuX7wRKpC4K5tnvlhgg/viewform Find the full list here: https://docs.google.com/spreadsheets/d/1HqTKukfJGpmowQnSh4CoFn3T6HXcNS1T1pK-Xx9CknQ/edit#gid=320641758
- This open letter from European scientists expressing solidarity and promoting peace is open for signatures: https://www.ipetitions.com/petition/an-open-letter-from-european-scientists-against
Additionally, I’m posting here some of the organizations responding to the crisis and actively soliciting donations:
- Doctors Without Borders / Médecins Sans Frontières (MSF): https://www.doctorswithoutborders.org/what-we-do/countries/ukraine
- The UN Refugee Agency: https://donate.unhcr.org/int/en/ukraine-emergency
- UNICEF: https://www.unicef.org/emergencies/conflict-ukraine-pose-immediate-threat-children
To help gather more support for these initiatives, please consider sharing this post further (you don’t need a ResearchGate account to see it), and I will continue to update it with other initiatives as I find them. You can also click “Recommend” below to help others in your ResearchGate network see it. And if you know of any other community initiatives that we can share here please let us know via this form: https://forms.gle/e37EHouWXFLyhYE8A
-Ijad Madisch, CEO & Co-Founder of ResearchGate
-----
Update 03/07:
This list outlines country-level initiatives from various academic institutions and research organizations, with a focus on programs and sponsorship for Ukrainian researchers:
Related Publications
The application of multiple regression and path analysis is discussed in regard to the exclusive use of the beta coefficients. Beta is one of the possible ways of controlling for the effects of the remaining predictors, others are part and partial correlation, part and partial covariance. A typology is developed for the difference between total and...
Path analysis was developed almost a century ago by geneticist Sewell Wright as a way to understand the relations among variables that goes beyond their correlations. While this method largely stayed in genetics and closely‐related fields, it eventually found its way into psychology in the later part of the twentieth century. This entry describes t...