What Are Acceptable Model Fit Standards?

What Are Acceptable Model Fit Standards?

Trần Huỳnh Gia HânHUF04 -

I used CFI and RMSEA to evaluate model fit. For CFI, I followed Hu and Bentler (1999), using ≥ .90 as acceptable and ≥ .95 as good fit. For RMSEA, based on Browne and ...

更多...

I used CFI and RMSEA to evaluate model fit. For CFI, I followed Hu and Bentler (1999), using ≥ .90 as acceptable and ≥ .95 as good fit. For RMSEA, based on Browne and Cudeck (1993), I used < .08 as acceptable and < .05 as good fit. My model met the acceptable thresholds (CFI > .90 and RMSEA < .08), but not the stricter “good fit” criteria. I would accept a model that “almost” fits, as long as it is theoretically sound, though improvements should still be considered.

What Are Acceptable Model Fit Standards?

Vo Dao Trang ThyHUF04 -
I agree with your thresholds because they are common and practical. I also think a model that “almost” fits can still be acceptable if it is theoretically meaningful. ...

更多...

I agree with your thresholds because they are common and practical. I also think a model that “almost” fits can still be acceptable if it is theoretically meaningful. However, I would still look at possible modifications before accepting it completely.

What Are Acceptable Model Fit Standards?

Hoàng Phan Trung HiếuHUF04 -
I find your choice of model fit thresholds quite strict, especially for RMSEA. While aiming for RMSEA < 0.05 is ideal, I think it may not always be realistic in applied ...

更多...

I find your choice of model fit thresholds quite strict, especially for RMSEA. While aiming for RMSEA < 0.05 is ideal, I think it may not always be realistic in applied research. I agree that context matters, and slightly higher values (e.g., up to 0.08) can still be acceptable if other indices show a good fit. Would you consider revising your model if it only “almost” meets the criteria, or would you accept it with justification?