The B-values represent standardized regression coefficients. Here's the rationale for them, put very simply:
- A t-test is great for answering the question "are two groups different" (e.g. "furries" vs. "non-furries"
- The problem is, in this study, we didn't have any "non-furries"
- So we did the next best thing: we used a continuous variable of "furriness": we asked participants to indicate how important furry was as a part of their identity on a 7-point scale. We used THIS as a proxy for "furry/non-furry"
- Now, we can't do a standard t-test on this data, because we don't have 2 different groups: there's all the people who said "1", plus those who said "2", plus those who said "3"... all the way to those who said "7". And it doesn't really make a lot of sense to treat them as 7 separate, completely independent groups: there IS a linear gradient (people who said "1" are a lot like people who said "2", and people who said "4" are probably more like people who said "5" than people who said "1" are.
- A linear regression basically plots the "line of best fit" for the data, and indicates what the slope of that line is (that's essentially what "B" is: the slope of that line). If it's positive, it means there's a positive relationship between "X" and "Y"; if it's negative, there's a negative relationship between "X" and "Y". The p-value afterward tells us if, statistically speaking, there's enough evidence to suggest that this slope is significantly different from "0" (which would mean "there is no relationship between X and Y"). If it's p<.05, then we say "yup, this relationship seems to be significant, because we wouldn't expect a slope that big to be due to simply to chance alone".
In sum, the B-value tells you the magnitude of the relationship between the two variables (and whether it's positive or negative), as imagined by the slope of a line that fits the data plot.
As for t-tests, a higher t-value tells you that the difference between the two groups is larger; it doesn't easily translate into something "practical"... Though, as a general rule of thumb, if the t-value is greater than about 2 (or smaller than -2), you can start to feel confident that the two groups are significantly different.
The B-values represent standardized regression coefficients. Here's the rationale for them, put very simply:
- A t-test is great for answering the question "are two groups different" (e.g. "furries" vs. "non-furries"
- The problem is, in this study, we didn't have any "non-furries"
- So we did the next best thing: we used a continuous variable of "furriness": we asked participants to indicate how important furry was as a part of their identity on a 7-point scale. We used THIS as a proxy for "furry/non-furry"
- Now, we can't do a standard t-test on this data, because we don't have 2 different groups: there's all the people who said "1", plus those who said "2", plus those who said "3"... all the way to those who said "7". And it doesn't really make a lot of sense to treat them as 7 separate, completely independent groups: there IS a linear gradient (people who said "1" are a lot like people who said "2", and people who said "4" are probably more like people who said "5" than people who said "1" are.
- A linear regression basically plots the "line of best fit" for the data, and indicates what the slope of that line is (that's essentially what "B" is: the slope of that line). If it's positive, it means there's a positive relationship between "X" and "Y"; if it's negative, there's a negative relationship between "X" and "Y". The p-value afterward tells us if, statistically speaking, there's enough evidence to suggest that this slope is significantly different from "0" (which would mean "there is no relationship between X and Y"). If it's p<.05, then we say "yup, this relationship seems to be significant, because we wouldn't expect a slope that big to be due to simply to chance alone".
In sum, the B-value tells you the magnitude of the relationship between the two variables (and whether it's positive or negative), as imagined by the slope of a line that fits the data plot.
As for t-tests, a higher t-value tells you that the difference between the two groups is larger; it doesn't easily translate into something "practical"... Though, as a general rule of thumb, if the t-value is greater than about 2 (or smaller than -2), you can start to feel confident that the two groups are significantly different.