24小时热门版块排行榜

返回列表

当前只显示满足指定条件的回帖，点击这里查看本话题的所有回帖

yuqimiao007

新虫 (小有名气)

应助: 0 (幼儿园)
金币: 414.5
帖子: 65
在线: 40.1小时
虫号: 1980024
注册: 2012-09-06
专业: 植物分类学

[求助] SMATR做标准化主轴分析(SMA)已有5人参与

刚开始接触学习标准化主轴分析(Standardized major axis regression, SMA)，遇到一些问题想请教大家：
1）SMA分析出来的斜率slope与excel中的斜率是否一致？我的数据计算出来是不一致的，但是我觉得理论上应该是一致的，不知道是不是我理解的不到位，在使用SMATR软件过程中出现错误？
2）两组或多组数据是否有共同斜率的P值判断。在"TEST FOR COMMON SLOPE ACROSS GROUPS" 这一栏里有个P值，在其后紧接着又有一个“Test statistic” 中也有一个P值，那么共同斜率的P值是看哪一个呢？后一个P值的意义是什么呢？
十分感谢大家能在百忙之中抽出时间来帮我解决问题。谢谢。
结果如下，请大家帮我分析一下，这两组数据是否有共同斜率(common slope)？谢谢。
SMA analysis
Filename: inputnp.txt
Grp: Group Y: log P X: log N
No. iterations: 1000 Critical p: 0.050 Conf.interval: 95%
Filters:
Filtered or empty data points: 16
Groups with less than 2 data points (removed):
No. groups: 2
No. datapoints: 251
Measurement Error: Not included

SMA results
Group    n       R2      p       Slope    LowCI    UppCI    Interc LowCI    UppCI    Ymean    Xmean    H0_b    F       p
1       133 0.567    0.000    1.431    1.278    1.603    -1.729 -1.937 -1.522 0.085    1.268    1.000    40.580 0.000
2       118 0.136    0.000    1.426    1.203    1.690    -1.801 -2.083 -1.519 -0.164 1.149    1.000    17.622 0.000

TEST FOR COMMON SLOPE ACROSS GROUPS
Grps    N       Slope    LowCI    UppCI    H0_b    X2      p
2       251 1.430    1.301    1.571    1.000    51.549 0.000

Test statistic: 0.001 p = 0.962

COMPARISON OF LINES WITH COMMON SLOPE
Common slope: 1.430 Grand mean X: 1.212 Grand mean Y: -0.032 Grand mean F: -1.764 Grand mean R: -1.764

Group n R2 Interc YgrandX XgrandY Ymean    Xmean    Fmean    Rmean
1       133 0.567 -1.727 0.005    1.186    0.085    1.268    1.897    -1.727
2       118 0.136 -1.806 -0.073 1.241    -0.164 1.149    1.478    -1.806

Testing for shifts in elevation between groups using WALD statistic
df stat p A_com
1 11.788 0.001    na
Significant elevation shift between groups!
No post-hoc multiple comparisons

Testing for shifts along the common slope using WALD statistic
df stat p A_com
1 99.678 0.000    na
Significant shift along common slope!
No post-hoc multiple comparisons

回复此楼