Community Forums

Share:

Chi-Square Exceeding Upper bounds Starnet  

Page 1 / 5

shibainafloat
Posts: 1
Member
(@shibainafloat)
FNG
Joined: 9 months ago

Hi All, 

I have a traverse and I am attempting to do a LSA on it with starnet. However, I am not sure why it is failing the chi-sq test. 

I determined a local grid and fixed coordinates for 2 points. Angles and distances were measured with a total station. 

Can anyone advise?

 

Here is my data: 

# local Grid

C E 5 48 !!

C D 5 75.977 !!

 

# Without crosshairs

TB B

T A 25-18-42.5 51.099

T E 155-53-30.5 27.977

T D 86-11-7 37.639

T C 66-54-40.5 34.91025

T B 205-42-26.5 40.961

TE A

# With crosshairs

M B-A-E 256-32-45.6 22.4735

M C-B-A 13-52-47 73.986

28 Replies
Bill93
Posts: 5567
Member
(@bill93)
5,000+ posts
Joined: 9 years ago

I haven't tried the traverse, but if doesn't have some obvious blunder in it, then anyone who wants to address the Chi-Square issue needs to know what standard errors you have given the program for angles, distances, instrument centering, and target centering.

One apparent problem is that you have fixed both coordinates of points D and E but then give traverse measurements through them.  Try freeing one of those four coordinate values.

Reply
Norman Oklahoma
Posts: 4501
Member
(@norman-oklahoma)
2,500+ posts
Joined: 4 years ago

Passes easily for me with the following standard errors:

Project Default Instrument
Distances (Constant) : 0.005000 FeetInt
Distances (PPM) : 2.000000
Angles : 3.000000 Seconds
Directions : 3.000000 Seconds
Azimuths & Bearings : 3.000000 Seconds
Centering Error Instrument : 0.003000 FeetInt
Centering Error Target : 0.010000 FeetInt

Reply
Dave Karoly
Posts: 10360
Member
(@dave-karoly)
10,000+ posts
Joined: 9 years ago

I took an Error Analysis Class.

The instructor said he doesn't even care if it passes Chi Square or not.  I got the vapors, they had to give me smelling salts.

If it exceeds the lower bound, I don't worry about it.  If it exceeds the upper bound then I look for my stupid mistake I can't see for the life of me even though it is staring right at me.

Reply
1 Reply
Gene Kooper
Member
(@gene-kooper)
Joined: 4 years ago

500+ posts
Posts: 998
Posted by: Dave Karoly

I took an Error Analysis Class.

The instructor said he doesn't even care if it passes Chi Square or not.  I got the vapors, they had to give me smelling salts.

If it exceeds the lower bound, I don't worry about it.  If it exceeds the upper bound then I look for my stupid mistake I can't see for the life of me even though it is staring right at me.

I agree with your instructor.  Look at the residuals.  If there is a blunder it will normally show up there.  If the residuals look good, who cares what a subjective "goodness of fit" statistical test says.  YMMV

Reply
thebionicman
Posts: 3174
Member
(@thebionicman)
2,500+ posts
Joined: 6 years ago

People who deal with social stats usually don't care about the chi square result. When you actually have to make things work it needs to pass...

Reply
Page 1 / 5
Default
:)
:d
:wink:
:mrgreen:
:neutral:
:twisted:
:arrow:
:shock:
:???:
:cool:
:evil:
:oops:
:razz:
:roll:
:cry:
:eek:
:lol:
:mad:
:sad:
:!:
:?:
:idea:
:hmm:
:beg:
:whew:
:chuckle:
:silly:
:envy:
:shutmouth: