Yes, there has been a discussion on this already. https://www.researchgate.net/post/In_a_multiple_regression_analysis_can_the_beta_coefficient_be_larger_than_1_and_if_so_is_there_something_wrong_in_the_analysis#:~:text=With%202%20or%20more%20predictors%20the%20betas%20can%20go%20beyond%20one.&text=Of%20course%20in%20multiple%20regression,your%20iv%20is%20in%20billion.
Yes they can be greater than 1 but 1.41 and 1.7 seem to be very odd in my opinion. Can you show your model and your results?
S. Béatrice Marianne Ewalds-Kvist are you sure about your answer in the context of linear regression? Although beta as volatility may be claculated with a multiple regression, the interpretastion is very specific here and does surely not generalize to beta coefficients in regression.
Or did you just google "beta coefficient" and copied and pasted a definition like from this homepage, which reads nearly identical:
https://www.investopedia.com/terms/b/beta.asp
Rainer Duesing I used linear regression by using excel and I used theory of planned behaviour model. When I calculated for whole sample, the value of beta coefficient was less than 1. But when I calculated for each faculty, the value of beta coefficient for two independence variables was greater than 1 for one faculty. In my opinion, it is because of small sample size (about 22).
Can you please show the output? And are you sure that you are talking about the standardized beta regression coefficient and not the unstandardized b coefficient? Unstandardized coefficients are in the scale of the dependent variable and can have any value on that scale. How many predictor variables?
And sample size shouldnt be an issue, since beta is just beta_x = b_x * sd(y)/sd(x).
Rainer Duesing Thanks for your explaining. This is the output from linear regression via excel. It is unstandardized coefficient. So do I need to calculate the standardized coefficient again?
For what do you need the standardized coefficients? Typically, it is much easier to interpret the unstandardized ones: a 1 unit change in the predictor variable equals a "b" unit change in the metric of the dependent variable. This is straight forward.
On the other hand, the standardized means a beta amount of change in the SD of the dependent variable for a change of 1 SD in the predictor variable. In my opinion, this is more ambigous than the original metric.
- Badges
- Science topic
- Similar topics
- Computer Science and Engineering
- Programming Languages
- Python