1.据我所知，r方不能完全判断模型的好坏，不然我们也不要r方了，直接看方差就行了。0.65左右我认为还行了，主要是看应用在什么方面。我至今还没弄懂这个pls到底是怎么算的，而且关于pls文献国内少之又少，要看就得看国外的。

2.mean=0，var=1 就是均值为0，方差为1，就是标准化，original就是非标，

3.才疏学浅，无法回答。

没人回复啊，我顶！

谢谢啊，我问了汉堡大学的那个老师了，他让我看看一篇文章来着。3个问题我在www.smartpls.de上面search了一下，他给了个链接。
Hi,

See TENENHAUS, M.; VINZI, V. E.; CHATELIN, Y.; LAURO, C. PLS path modeling. Computational Statistics & Data Analysis, v.48, p.159-205, 2005.
Available at: https://studies2.hec.fr/jahia/we ... ticles/PLS_PM_5.pdf

In this article you will find a GoF index that is computed like a geometric mean between structural (R2) and measurement (AVE) model (reflective LV), but no threshold was proposed.

Remembering that is recommended that:
AVE > 0.5

R2 depends on the number of predictors, area of research, as Cohen (1977, p.79) had said:
“A reader who finds what is here defined as ‘large’ is too small (or too large) to meet what his area of behavioral science would consider appropriate standards is urged to make more suitable operational definitions. What are offered below are definitions for use when no others suggest themselves, or as conventions.”

Simple correlations (p.79-80):
Small effect = 0.10
Medium effect = 0.30
Large effect = 0.5

And for multiple regression (or PLS-PM) we could use the recommendation in the page 413.
Efeito pequeno f2 = 0,02 --> R2 = 2%
Efeito médio f2 = 0,15 --> R2 = 13%
Efeito grande f2 = 0,35 --> R2 = 26%


COHEN, J. Statistical Power Analysis for the Behavioral Sciences. Revised Edition. New York: Academic Press, 1977.


More answers in the article.

Best regards,

Bido

关于第三个问题，我在网站上找的的结果，供大家参考（不知道括号里面对不对，我自己理解写的）：
cases = number of cases (or observations) in your colected sample （观测变量的数量）
sample = number of times that new samples will be re-sampled from the raw data and the parameters of the model will be computed - usually 500 or 1000（计算的次数）

我的文章是写客户关系价值的，是个3阶的SEM。
我看台湾同胞用PLS衡量的时候，涉及T值，CR和AVE，R2也提及了。但是由于是PPT，没写太详细的东西。
老外Wolfgang的文章用的也是PLS，收敛效度涉及了T值，a值，AVE和负载loading，区分效度用了AVE和协相关关系方差，EFA的数据给了一点，再就没有其他的了。发表在JM上。没有R2和CR的东西。