Jump to content
House Price Crash Forum

Archived

This topic is now archived and is closed to further replies.

Smell the Fear

I Need A Statistician! Help!

Recommended Posts

Hi All

I haven't been posting here for a while due to other commitments, but rest assured I am feeling the same combination of glee at being right and horror at the unfolding economic disaster that most of you are (no jokes about my "bull" phase, it was a joke!).

As this place is the finest repository of intellectual muscle in the known world I thought I should post this request here. I need someone with a sound knowledge of statistics to interpret some summary data for me. This is extremely important and I will be eternally grateful - you could potentially save my career and the positive karma will see you through the recession/depression/wipeout that we are entering.

Here goes with the data:

t-Test: Two-Sample Assuming Equal Variances

Group 1 Group 2

Mean 70.35 70.50

Variance 36.83 32.35

Observations 159 87

Pooled Variance 35.25

Hypothesized Mean Difference 0

df 244

t Stat -0.188

P(T<=t) one-tail 0.425

t Critical one-tail 1.65

P(T<=t) two-tail 0.85

t Critical two-tail 1.96

F-Test Two-Sample for Variances

Overall Comparison Group 1 Group 2

Mean 70.35 70.5

Variance 36.83 32.35

Observations 159 87

df 158 86

F 1.138

P(F<=f) one-tail 0.255

F Critical one-tail 1.379

Please, only repond if you have a firm grip of statistics. I don't mean to be rude, it's just that I need some reliable information.

Thanks

STF

Share this post


Link to post
Share on other sites
Guest DissipatedYouthIsValuable

I'll do it. But only if you admit to being a weedy ponce retard who was a bull.

Share this post


Link to post
Share on other sites

Bump!

Some of the financial gurus must know what this means. I think I have a broad grasp but I need clear, concise and logical explanations from a stats master.

please.......... :(

Share this post


Link to post
Share on other sites
Hi All

I haven't been posting here for a while due to other commitments, but rest assured I am feeling the same combination of glee at being right and horror at the unfolding economic disaster that most of you are (no jokes about my "bull" phase, it was a joke!).

As this place is the finest repository of intellectual muscle in the known world I thought I should post this request here. I need someone with a sound knowledge of statistics to interpret some summary data for me. This is extremely important and I will be eternally grateful - you could potentially save my career and the positive karma will see you through the recession/depression/wipeout that we are entering.

Here goes with the data:

t-Test: Two-Sample Assuming Equal Variances

Group 1 Group 2

Mean 70.35 70.50

Variance 36.83 32.35

Observations 159 87

Pooled Variance 35.25

Hypothesized Mean Difference 0

df 244

t Stat -0.188

P(T<=t) one-tail 0.425

t Critical one-tail 1.65

P(T<=t) two-tail 0.85

t Critical two-tail 1.96

F-Test Two-Sample for Variances

Overall Comparison Group 1 Group 2

Mean 70.35 70.5

Variance 36.83 32.35

Observations 159 87

df 158 86

F 1.138

P(F<=f) one-tail 0.255

F Critical one-tail 1.379

Please, only repond if you have a firm grip of statistics. I don't mean to be rude, it's just that I need some reliable information.

Thanks

STF

er, you havent actually asked a question have you? "interpret the data" means nothing un;ess you tell us what the data is measuring. you may as well say "is this piece of paper big enough?"

Share this post


Link to post
Share on other sites

I will have a look at it later - a bit busy now. When do you need the answer?

I have a PhD in statistics. I assume you want to know whether the means of the two samples are different at a statitically significant level. In other words, I assume your null hypothesis is that Mean of Sample A = Mean of Sample B and that you are therefore looking to reject the null at some degree of staistical confidence?

WARNING: I wil have to check my answer in a text book as it is a few years since I did this particular calculation. In fact I think you could do it on Excel with one of the functions there but cannot remember which.

Share this post


Link to post
Share on other sites
Hi All

I haven't been posting here for a while due to other commitments, but rest assured I am feeling the same combination of glee at being right and horror at the unfolding economic disaster that most of you are (no jokes about my "bull" phase, it was a joke!).

(snip)

Thanks

STF

Have a read of this:

http://en.wikipedia.org/wiki/Student%27s_t-test

http://en.wikipedia.org/wiki/F_test

This should help you make sense of the above.

Share this post


Link to post
Share on other sites
Hi All

I haven't been posting here for a while due to other commitments, but rest assured I am feeling the same combination of glee at being right and horror at the unfolding economic disaster that most of you are (no jokes about my "bull" phase, it was a joke!).

As this place is the finest repository of intellectual muscle in the known world I thought I should post this request here. I need someone with a sound knowledge of statistics to interpret some summary data for me. This is extremely important and I will be eternally grateful - you could potentially save my career and the positive karma will see you through the recession/depression/wipeout that we are entering.

Here goes with the data:

t-Test: Two-Sample Assuming Equal Variances

Group 1 Group 2

Mean 70.35 70.50

Variance 36.83 32.35

Observations 159 87

Pooled Variance 35.25

Hypothesized Mean Difference 0

df 244

t Stat -0.188

P(T<=t) one-tail 0.425

t Critical one-tail 1.65

P(T<=t) two-tail 0.85

t Critical two-tail 1.96

F-Test Two-Sample for Variances

Overall Comparison Group 1 Group 2

Mean 70.35 70.5

Variance 36.83 32.35

Observations 159 87

df 158 86

F 1.138

P(F<=f) one-tail 0.255

F Critical one-tail 1.379

Please, only repond if you have a firm grip of statistics. I don't mean to be rude, it's just that I need some reliable information.

Thanks

STF

This looks like the cut & pasted output of the microsoft excel analysis toolpack to me. Without knowing what the data is supposed to represent, I can't comment on what it might mean though.

Share this post


Link to post
Share on other sites
I will have a look at it later - a bit busy now. When do you need the answer?

I have a PhD in statistics. I assume you want to know whether the means of the two samples are different at a statitically significant level. In other words, I assume your null hypothesis is that Mean of Sample A = Mean of Sample B and that you are therefore looking to reject the null at some degree of staistical confidence?

WARNING: I wil have to check my answer in a text book as it is a few years since I did this particular calculation. In fact I think you could do it on Excel with one of the functions there but cannot remember which.

:) analysis tool pack.

Share this post


Link to post
Share on other sites
I will have a look at it later - a bit busy now. When do you need the answer?

I have a PhD in statistics. I assume you want to know whether the means of the two samples are different at a statitically significant level. In other words, I assume your null hypothesis is that Mean of Sample A = Mean of Sample B and that you are therefore looking to reject the null at some degree of staistical confidence?

WARNING: I wil have to check my answer in a text book as it is a few years since I did this particular calculation. In fact I think you could do it on Excel with one of the functions there but cannot remember which.

Great, thanks!

I think the answer to the question should be in the information I originally posted. It is the interpretation of the information that I need help with. My understanding so far is as follows:

What we have are two groups with values assigned to them dependent upon their performance. Each group is made of a number of individuals as shown by the number of observations. Each group has a mean, and a variance.

The t-test looks at the significance of the difference in their means I think. I think the one tail P value means that there is a 42% chance that the difference in the means is due to chance, hence a 58% chance that it is not due to chance. Is this the case? Would it be true to say that on a balance of probabilities it is likely that the difference is not due to chance?

I don't know what the difference is between the one tail and two tail or what "t critical one tail" means.

I think the F test is similar but applied to the variance of the data rather than the means.

Is it the case that "statistical significance" is simply an artificial construct in statistics which is generally used to signify an overwhelming weight of evidence in favour of a hypothesis?

Share this post


Link to post
Share on other sites
Have a read of this:

http://en.wikipedia.org/wiki/Student%27s_t-test

http://en.wikipedia.org/wiki/F_test

This should help you make sense of the above.

Already have read it, good start but I need an expert who can lay down a correct interpretation quickly - it would take weeks or months to get to grip with this fully and I would have no way of telling if I was making fatal errors.

Share this post


Link to post
Share on other sites
Already have read it, good start but I need an expert who can lay down a correct interpretation quickly - it would take weeks or months to get to grip with this fully and I would have no way of telling if I was making fatal errors.

In that case you'd need to be taught to degree level to understand it, and unless you want to be a stats pro, what's the point of that?

Why don't you pay a professional to do this for you? If money is riding on a correct interpretation, you'd want the data checked as well really -- it's weird that there is a calculation but no interpretation, so I'd be loathe to trust the math.

Share this post


Link to post
Share on other sites
In that case you'd need to be taught to degree level to understand it, and unless you want to be a stats pro, what's the point of that?

Why don't you pay a professional to do this for you? If money is riding on a correct interpretation, you'd want the data checked as well really -- it's weird that there is a calculation but no interpretation, so I'd be loathe to trust the math.

I'm on my own I'm afraid, can't afford to pay an expert and don't have access to the raw data. Life isn't always fair, I'm trying to do the best with what I have available.

Share this post


Link to post
Share on other sites

My comments in caps below - hope this helps.

Hi All

I haven't been posting here for a while due to other commitments, but rest assured I am feeling the same combination of glee at being right and horror at the unfolding economic disaster that most of you are (no jokes about my "bull" phase, it was a joke!).

As this place is the finest repository of intellectual muscle in the known world I thought I should post this request here. I need someone with a sound knowledge of statistics to interpret some summary data for me. This is extremely important and I will be eternally grateful - you could potentially save my career and the positive karma will see you through the recession/depression/wipeout that we are entering.

Here goes with the data:

t-Test: Two-Sample Assuming Equal Variances

Group 1 Group 2

Mean 70.35 70.50

Variance 36.83 32.35

Observations 159 87

WE HAVE TWO SETS OF DATA, GROUP 1 AND GROUP 2

EACH GROUP HAS A SET OF STATISTICS ASSOCIATED WITH IT, AND A CERTIAN NUMBER OF DATA POINTS

MOST OBVIOUS SENSIBLE QUESTION WHICH THIS ANALYSIS COULD BE ADDRESSING IS WHETHER THE UNDERLYING DISTRIBUTIONS OF THESE GROUPS ARE THE SAME, OR MORE PRECISELY, WHETHER THIS DATA ALLOWS US TO CONCLUDE WITH ANY LEVEL OF CERTAINTY THAT THEY ARE DIFFERENT.

Pooled Variance 35.25

FIRST WE TEST USING THE T-TEST WHETHER THE MEANS OF THE GROUPS ARE SIGNIFICANTLY DIFFERENT

Hypothesized Mean Difference 0

df 244

t Stat -0.188

THE T-STATISTIC IS CALCULATED FROM THE DATA TO APPLY THE T-TEST

P(T<=t) one-tail 0.425

t Critical one-tail 1.65

P(T<=t) two-tail 0.85

t Critical two-tail 1.96

THESE RESULTS ESSENTIAL STATE THAT, GIVEN THE VALUE FOR THE T-STATISTIC THERE IS NO EVIDENCE AT ALL THAT THE MEANS ARE SIGNIFICANTLY DIFFERENCE (THE T-STAT WOULD HAVE TO BE OUTSIDE [-1.65,1.65] FOR THAT TO BE THE CASE)

F-Test Two-Sample for Variances

SECONDLY WE USE THE F-TEST TO SEE IF THE VARIANCES ARE SIGNIFICANTLY DIFFERENT

Overall Comparison Group 1 Group 2

Mean 70.35 70.5

Variance 36.83 32.35

Observations 159 87

df 158 86

F 1.138

WE CALCULATE THE F-STATISTIC IN ORDER TO APPLY THE F-TEST

P(F<=f) one-tail 0.255

F Critical one-tail 1.379

THE F-STATISTIC INDICATES THAT THERE IS NOT A STATISTICALLY SIGNIFICANT DIFFERENCE BETWEEN THE VARIANCES OF THE TWO GROUPS, ALTHOUGH COMPARING THE 1.138 AND THE 1.379 SUGGESTS THAT IT IS RELATIVELY CLOSE TO BEING SIGNIFICANT.

Please, only repond if you have a firm grip of statistics. I don't mean to be rude, it's just that I need some reliable information.

Thanks

STF

Share this post


Link to post
Share on other sites
I'm on my own I'm afraid, can't afford to pay an expert and don't have access to the raw data. Life isn't always fair, I'm trying to do the best with what I have available.

What are you measuring and how is it measured?

Unless you can tell us what it is, or at least confirm that it is a continuous variable, we can't tell you have performed the correct analysis.

Indeed, we don't have enough information. Were other parameters measured, other than just the one for which mean/variance are calculated? If they were, and there is possibility of confounding, a multivariable analysis may be better.

Share this post


Link to post
Share on other sites
What are you measuring and how is it measured?

Unless you can tell us what it is, or at least confirm that it is a continuous variable, we can't tell you have performed the correct analysis.

Indeed, we don't have enough information. Were other parameters measured, other than just the one for which mean/variance are calculated? If they were, and there is possibility of confounding, a multivariable analysis may be better.

exactly, if youre measuring the peformance of 2 groups of aircraft bolts (1 you already use and 1 which is cheaper and youre thinking of using) you will have a different interpretation of data. ie youd want to be almost 100% certain the 2 group are not statistically different.

as opposed to 2 groups of scarf lengths where you really dont care too much as long as theyre in the same ballpark

Share this post


Link to post
Share on other sites
Great, thanks!

I

I don't know what the difference is between the one tail and two tail or what "t critical one tail" means.

one tail h0 is equal to 100

h1 is not equal to 100

2 tail test ho is greater or equal to 100

h1 is less than 100.

id t or F is greater than p we reject ho and accept h1

f test f value is less than or equal to the critical value we accept ho, if greater we reject then we accept h1.

so if P is lower than significance level then we reject ho

if p is greater than significance level we accept ho and reject h1.

it is a way to find out if h0 is false or true.

Share this post


Link to post
Share on other sites
My comments in caps below - hope this helps.

Agreed! The only thing I would add is the difference between the 1 tail and 2 tail test: the one tail is testing that, on average, group 2 is greater than group 1, the two tail just that they are different. I suspect you don't know it, but you've tested at the 95% significance level which is the conventional level used by stattos, and thus usually set as default in programs.

One final thing to be careful of is your assumptions, these are detailed at the top of the link given by Cinnamon. In particular, you need to be careful that your observations are independent, as if they're not, your conclusions / models could be wrong. I understand this was a key mistake made by the boffins whilst doing their mortgage / loan risk analysis: they forgot that loan defaults are not independent (look, I'm on topic :lol: )

Share this post


Link to post
Share on other sites
Agreed! The only thing I would add is the difference between the 1 tail and 2 tail test: the one tail is testing that, on average, group 2 is greater than group 1, the two tail just that they are different. I suspect you don't know it, but you've tested at the 95% significance level which is the conventional level used by stattos, and thus usually set as default in programs.

One final thing to be careful of is your assumptions, these are detailed at the top of the link given by Cinnamon. In particular, you need to be careful that your observations are independent, as if they're not, your conclusions / models could be wrong. I understand this was a key mistake made by the boffins whilst doing their mortgage / loan risk analysis: they forgot that loan defaults are not independent (look, I'm on topic :lol: )

Daedalus/Datamonkey - thanks very much. That confirms what I thought, more or less.

When you say "there is no evidence at all that the means are different" is that a fair comment? Is it not the case that a P value of say 0.25 would indicate a 75% probability that the difference was due to factors other than chance? In scientific terms that would be rejected but in other areas of life it may be regarded as sufficient. Please correct me if this is wrong as I don't want to look like an idiot by stating it! :(

Share this post


Link to post
Share on other sites

I could have helped you about 10 years ago, but I've forgotten most of that now. But it seems some of the other guys remember it.

But you don't seem to have said what information or statistic you are looking for from the data and stats you have provided.

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.

  • 396 Brexit, House prices and Summer 2020

    1. 1. Including the effects Brexit, where do you think average UK house prices will be relative to now in June 2020?


      • down 5% +
      • down 2.5%
      • Even
      • up 2.5%
      • up 5%



×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.