Warning: Illegal string offset 'html' in /home/hsn/public_html/forum/cache/skin_cache/cacheid_1/skin_topic.php on line 909

Statistics question - HSN forum

# Statistics question

5 replies to this topic

### #1ferg

Good Effort

• Members
• 64 posts
• Location:Stranraer
• Interests:driving, music, architecture
• Gender:Male

Posted 06 April 2006 - 12:43 PM

in a chi-square calc i have been asked to use both significance levels. in this case the significance gives two critical values 1 smaller than the calculated value and 1 larger. this means that the null hypothesis could be accpeted but also could be rejected.

what is the reason for the different significance levels giving different answers (ie reject/accept) is it something to do with chance?

all help appreciated

### #2bred

Brendan

• Moderators
• 4,215 posts
• Location:Edinburgh
• Interests:I have just graduated with a 2:1 in Geography [MA (Hons)] from The University of Edinburgh. I like sports: swimming, cycling, snowboarding, running, football, mountain biking and also travelling and photography.
• Gender:Male

Posted 06 April 2006 - 12:56 PM

The significance levels give an indication of whether there is a correlation or not. At 99% significance you can say that there is almost certainly a correlation. 95% significance also indicates a correlation, but not as strongly as 99%.

If your answer is below the critical value necessary for 99% chance of significance but above the critical value necessary for 95% chance of significance then you can reject the null hypothesis and say there appears to be a correlation, albeit not a very strong one.
Please vote for me! (Brendan Howard, 5th from bottom, only 1 vote required): http://answers.polld...m/poll/2330393/

### #3ferg

Good Effort

• Members
• 64 posts
• Location:Stranraer
• Interests:driving, music, architecture
• Gender:Male

Posted 06 April 2006 - 01:02 PM

thnx! always knew a could rely on hsn 4 an answer!

### #4ferg

Good Effort

• Members
• 64 posts
• Location:Stranraer
• Interests:driving, music, architecture
• Gender:Male

Posted 06 April 2006 - 01:51 PM

next problem!... with nearest neighbour analysis

my answer for Rn is 1.52 which is pretty much in the middle of Regular and Random, what does it go with or does it show that some of the villages are in a regular patern while the others are random??

### #5bred

Brendan

• Moderators
• 4,215 posts
• Location:Edinburgh
• Interests:I have just graduated with a 2:1 in Geography [MA (Hons)] from The University of Edinburgh. I like sports: swimming, cycling, snowboarding, running, football, mountain biking and also travelling and photography.
• Gender:Male

Posted 16 April 2006 - 12:39 PM

That result indicates a fairly regular pattern, enough to be considered statistically significant:

Results outwith the yellow shaded area can be taken as statistically significant.

As you can see above, the Rn value needs to be quite close to 1 for it to be insignificant either way (especially if there are many values).
Please vote for me! (Brendan Howard, 5th from bottom, only 1 vote required): http://answers.polld...m/poll/2330393/

### #6ferg

Good Effort

• Members
• 64 posts
• Location:Stranraer
• Interests:driving, music, architecture
• Gender:Male

Posted 17 April 2006 - 02:52 PM

thnx again! im the only person doing ah geo @ my school, so when i get stuck i have no 1 2 get help from!

#### 1 user(s) are reading this topic

0 members, 1 guests, 0 anonymous users