| Original message ([an error occurred while processing this directive] Views )[an error occurred while processing this directive]
|
Replies:
|
| Makondo99
128th Post
 
Regular Customer
   
|  "The terms"
, posted Tue 7 Jan 08:14
quote:
Graphic design's not much, but hope it helps regardless...
Whoa! That's of great help! Thanks OmegaDog!
Anyway, I'd be quite grateful if any caritative soul can give me a hand with this:
I'll leave in bold those terms that are giving me a headache. Please excuse my lousy translation.
First thing it's to tackle this figure:
http://makondo99.tripod.com/stuff/table_1.gif
Here I go:
Clase: refers to the range of scores of the test (from 0 to 5 in the writing test and from 1 to 12 in the reading test).
Frecuencia: refers to the amount in which each clase is repeated in the sample chosen.
Frecuencia relativa: refers to the percentage that each clase stands for. (fi =(ni*100)/N).
N: refers to figure obtained by adding all the frecuencias.
F: refers to the percentage obtained by adding all the frecuencias relativas, i.e. 100%.
I need a blanket term to refer to the range that each frecuencia has in the chart. A loose approximation would be the spread of frecuencias, but Im not sure if it makes sense in English.
This is an actual chart that I have to analyse:
http://makondo99.tripod.com/stuff/Chart.gif
Secondly, I have to refer to crazy terms I'd never seen before. You can laugh at this poor translation (I bet even Babelfish can do a better job...):
Crappy translation: (I'm refering to this http://makondo99.tripod.com/stuff/TStudent.gif graphic)
prueba t de Student (Uh, Students t test ?), it measures whether there is a significant difference in the results obtained by the application of two different approaches on a sample
The analysis consists in obtaining a ts value (Im not sure if value is a correct translation) that corresponds to a valor muestral (sample value?) obtained from the analysis of the data collected and compare this value (?) against a standard t value (?) taken from a statistical table, in which you can find that for a specific number of observations (?) a determined quotient called grados de libertad (Uh, grade of looseness/freedom?, it doesnt make much sense, does it?) and a determined nivel de confianza (trust level?? %?).
ts is obtained using this formula:
ts = ((X1-X2)/s )*(((N1*N2)/(N1+N2))^0.5)
Where:
X1: corresponds to the average score obtained by the control group.
X2: corresponds to the average score obtained by the experimental group.
s: corresponds to the standard deviation.
N1: is the figure obtained by adding up all the scores of the control group.
N2: is the figure obtained by adding up all the scores of the experimental group
The grados de libertad are calculated using the following formula:
GL = N 1
Where:
N: corresponds to the whole number of subjects who compose a group.
A nivel de confianza of 99% has been chosen arbitrarily, therefore the GL in the case of men is 25 (the whole group are actually 26 men) and 19 in the case of women (accordingly, the whole group of women are 20 people).
So, for men t(99%) = 2,48 y and for women t(99%) = 2,53.
Help!
|
| Gen 1149th Post
Red Carpet Premium Member+
 
    
    
| "Re(1):Help with statistics" , posted Tue 7 Jan 12:09:
I'm writing out statistics information to remind my self of instead of a line by line response to your question.
Population is the group from which the statistics are taken, such as males age 25-30 at a Brown university.
Sample is a sub set of population. A sample of the population is tested because it's more scientific than measuring everyone and more practical.
The sample group should be chosen randomly or by scientific selection.
Mean average is the regular average. Adding all the figures and dividing by the number of figures gives the mean average.
The Mu symbol can be used for the mean, I don't know the code for typing it off hand right now. It looks like a "u" with some extra lines.
The mean can also be written as an x with a horizontal bar over it. I will use "x bar" in place of the symbol.
An angular letter "E" looking symbol, it might be sigma, means sum of. I will use "E" in place of the Greek letter.
We'll let n=the number of numbers in our group. So the average can be written as:
x bar = Ex
x bar = ___
x bar = n
So if your numbers are 1,4,5,5,5,6,7,8,9,9,10,11
then the sum of x or Ex = 79
There are 12 (edit: oops I had 11 here before but it's 12) numbers listed so n = 12
x bar = Ex spacer x bar = 80
x bar = ___ spacex bar = ___
x bar = n spacerrrx bar = 12
x bar = 5
Mode average is the most frequent number. So if your numbers are 1,4,5,5,5,6,7,8,9,9,10,11
5 appears 3 times. 9 appears 2 times. The rest appear one time. Sense 5 appears the most it is the mode average.
Median average I think is the center of the list. If there are an even amount of numbers listed, and the two in the middle are different, than either both can be listed or both can then be mean averaged. So if this is correct, the median of 1,4,5,5,5,6,7,8,9,9,10,11 would be 6 and/or 7, or the mean average of the two which is 6.5. If you have an odd amount of numbers, like 1,5,5,5,6,7,8,9,9,10,11; then the median is just the one in the middle which is 7.
One word for a division of the range of numbers is interval. So we could have the intervals 1-3, 4-6, 7-9, and 10-12. You could probably also call them ranges or some other names.
for 1,4,5,5,5,6,7,8,9,9,10,11
we have
1#################### in# 1- 3 interval
##4,5,5,5,6############# in# 4- 6
######### 7,8,9,9####### in# 7- 9
############### 10,11 # in 10-12
counting them we have
1 in# 1-3
5 in# 4-6
4 in# 7-9
2 in 10-12
later we might use some different intervals.
Differences from the mean or x - x bar
This could also be called deviation from the mean.
#1 - 5 = -4
#4 - 5 = -1
#5 - 5 = # 0
#5 - 5 = # 0
#5 - 5 = # 0
#6 - 5 = # 1
#7 - 5 = # 2
#8 - 5 = # 3
#9 - 5 = # 4
#9 - 5 = # 4
#9 - 5 = # 4
10 - 5 = # 5
11 - 5 = # 6
I will use "^2" to mean square.
Now lets do deviation from the mean squared
(x - x bar)^2.
####### x - x bar ##### (x - x bar)^2
#1 - 5 = -4 ######### 16
#4 - 5 = -1########## 1
#5 - 5 =# 0########## 0
#5 - 5 =# 0########## 0
#5 - 5 =# 0########## 0
#6 - 5 =# 1########## 1
#7 - 5 =# 2########## 4
#8 - 5 =# 3########## 9
#9 - 5 =# 4######### 16
#9 - 5 =# 4######### 16
#9 - 5 =# 4######### 16
10 - 5 =# 5######### 25
11 - 5 =# 6######### 36
Now we'll calculate the totals, mainly the Sum of the squares of the difference
E(x - x bar)^2
####### x - x bar ##### (x - x bar)^2
#1 - 5 = -4 ######### 16
#4 - 5 = -1########## 1
#5 - 5 =# 0########## 0
#5 - 5 =# 0########## 0
#5 - 5 =# 0########## 0
#6 - 5 =# 1########## 1
#7 - 5 =# 2########## 4
#8 - 5 =# 3########## 9
#9 - 5 =# 4######### 16
#9 - 5 =# 4######### 16
#9 - 5 =# 4######### 16
10 - 5 =# 5######### 25
11 - 5 =# 6######### 36
Totals:
80####24#########140
Ex = 80
E(x - x bar) = 24
E(x - x bar)^2 = 140
Note that the Sum of the Squares of (x - x bar) is written like this
E(x - x bar)^2
Which is different than the Square of the Sum of (x - x bar) which could be written:
(E(x - x bar))^2
Variance = E(x - x bar)^2
#########______________
#########n - 1
Variance = 140
#########______________
#########12-1
Variance = 140
#########______________
######### 11
Variance = 12.7272727272727272727272727272727...
Standard Deviation from the mean is the square root of the variance.
Standard Deviation =
_\\///________________
_\\/// E(x - x bar)^2
_\\/// ______________
_\\/// n - 1
To be certain of the terms I'd buy an English language college statistics text book, meant for a course the only purpose of which is to teach statistics instead of statistics tacked on to some other discipline.
[this message was edited by Gen on Tue 7 Jan 17:14] |
| Gen 1149th Post
Red Carpet Premium Member+
| "Re(1):The terms" , posted Tue 7 Jan 13:13:
First thing it's to tackle this figure:
http://makondo99.tripod.com/stuff/table_1.gif
Here I go:
Clase: refers to the range of scores of the test (from 0 to 5 in the writing test and from 1 to 12 in the reading test).
This is probably interval or range.
Frecuencia: refers to the amount in which each clase is repeated in the sample chosen.
Probably frequency.
Frecuencia relativa: refers to the percentage that each clase stands for. (fi =(ni*100)/N).
Probably relative frequency.
N: refers to figure obtained by adding all the frecuencias.
F: refers to the percentage obtained by adding all the frecuencias relativas, i.e. 100%.
I need a blanket term to refer to the range that each frecuencia has in the chart. A loose approximation would be the spread of frecuencias, but Im not sure if it makes sense in English.
Probably range or interval again? Or range of the interval. Or interval range. I also hear spread in sports statistic alot as in "can they beat the spread?" I think if you have the numbers 4,8,9,10. Then the "spread" would be 6.
This is an actual chart that I have to analyse:
http://makondo99.tripod.com/stuff/Chart.gif
Secondly, I have to refer to crazy terms I'd never seen before. You can laugh at this poor translation (I bet even Babelfish can do a better job...):
Crappy translation: (I'm refering to this http://makondo99.tripod.com/stuff/TStudent.gif graphic)
prueba t de Student (Uh, Students t test ?), it measures whether there is a significant difference in the results obtained by the application of two different approaches on a sample
You don't mean the T-Score?
The analysis consists in obtaining a ts value (Im not sure if value is a correct translation) that corresponds to a valor muestral (sample value?) obtained from the analysis of the data collected and compare this value (?) against a standard t value (?) taken from a statistical table, in which you can find that for a specific number of observations (?) a determined quotient called grados de libertad (Uh, grade of looseness/freedom?, it doesnt make much sense, does it?) and a determined nivel de confianza (trust level?? %?).
maybe you do mean T-score. You have a t-score chart right (no one memorizes them anymore)?
grados de libertad - in English we say "Degrees of freedom" so that's probably it
nivel de confianza - in statistics we talk about confidence. This would be confidence level. You can decide how "confident" the results will be, statistics often uses 95%, 99%, 99.5%... You could say "99.5% confidence".
ts is obtained using this formula:
ts = ((X1-X2)/s )*(((N1*N2)/(N1+N2))^0.5)
Where:
X1: corresponds to the average score obtained by the control group.
X2: corresponds to the average score obtained by the experimental group.
s: corresponds to the standard deviation.
N1: is the figure obtained by adding up all the scores of the control group.
N2: is the figure obtained by adding up all the scores of the experimental group
The grados de libertad are calculated using the following formula:
GL = N 1
Where:
N: corresponds to the whole number of subjects who compose a group.
A nivel de confianza of 99% has been chosen arbitrarily, therefore the GL in the case of men is 25 (the whole group are actually 26 men) and 19 in the case of women (accordingly, the whole group of women are 20 people).
So, for men t(99%) = 2,48 y and for women t(99%) = 2,53.
Help!
I have a kick ass text book let me see if I can find the name of it...
[this message was edited by Gen on Tue 7 Jan 17:20] |
| Makondo99
129th Post
Regular Customer
| "T-Score?" , posted Tue 7 Jan 16:59
THANKS A LOT for the feedback folks, I updated and corrected the text accordingly.
It's a long text, so you may skip it to the only part that it's still bugging me: The T-Score
***Updated Text***
Explaining this chart:
http://makondo99.tripod.com/stuff/table_1.gif
Sex?: indicates to which sex corresponds the group (male or female)
Stage?: corresponds to stage to which the data belongs to (pre- or post-test stage)
Group?: corresponds to the kind of group (experimental or control)
Test?: corresponds to the kind of test (reading or writing)
Range: corresponds to the possible range of scores that could be reached in the test (from 0 to 5 in the writing test and from 1 to 12 in the reading test)
Frequency: indicates how often a range occurs
Relative Frequency: reflects in percentage the frequency of each range using the formula fi = (ni*100)/N)
N: refers to the figure obtained by adding all the frequencies up
F: refers to the percentage obtained by adding all the relative frequencies up, i.e. F = 100%
***end***
*** Text I'm still unsure of***
Validating the graphic analysis:
In order to validate the previous analysis, a t-test was applied. It consisted in calculating a t-score (ts) and then comparing it to a standard t-value taken from a t-score chart (T).
The t-score was computed using the following formula:
ts = ((X1-X2)/s )*(((N1*N2)/(N1+N2))^0.5)
Where:
X1: corresponds to the mean of the control group.
X2: corresponds to the mean of the experimental group.
s: corresponds to the standard deviation.
N1: is the figure obtained by adding up all the scores of the control group.
N2: is the figure obtained by adding up all the scores of the experimental group
The standard deviation s was calculated using the corresponding formula in the Excel program. The range used was a column d that contains the results of the subtraction of the scores obtained by the control group minus the scores obtained by the experimental group.
In order to pick up a t-value from the t-chart, a degree of freedom has to be calculated beforehand using the following formula:
DF = N 1
Where N corresponds to the group with the highest numbers of subjects that is being compared with the other group.
Therefore the DF in the case of the male group is 26 and 19 in the case of the female group. A confidence level of 99% was then chosen that corresponds to the followings values in the t-chart (T): T(0.99) = 2,4786 for males and T(0.99) = 2,5395 for females.
The interpretation of the t-score is done according to these criteria:
If ts>T, then there is a significant difference between the groups.
If ts<t, then there is not any significant difference between the groups.
Am I correct in THAT is what the T-Score measure?
|
| Radish 730th Post
Red Carpet Regular Member+
| "Re(1):T-Score?" , posted Tue 7 Jan 19:34
quote:
THANKS A LOT for the feedback folks, I updated and corrected the text accordingly.
It's a long text, so you may skip it to the only part that it's still bugging me: The T-Score
***Updated Text***
Explaining this chart:
http://makondo99.tripod.com/stuff/table_1.gif
Sex?: indicates to which sex corresponds the group (male or female)
Stage?: corresponds to stage to which the data belongs to (pre- or post-test stage)
Group?: corresponds to the kind of group (experimental or control)
Test?: corresponds to the kind of test (reading or writing)
Range: corresponds to the possible range of scores that could be reached in the test (from 0 to 5 in the writing test and from 1 to 12 in the reading test)
Frequency: indicates how often a range occurs
Relative Frequency: reflects in percentage the frequency of each range using the formula fi = (ni*100)/N)
N: refers to the figure obtained by adding all the frequencies up
F: refers to the percentage obtained by adding all the relative frequencies up, i.e. F = 100%
***end***
*** Text I'm still unsure of***
Validating the graphic analysis:
In order to validate the previous analysis, a t-test was applied. It consisted in calculating a t-score (ts) and then comparing it to a standard t-value taken from a t-score chart (T).
The t-score was computed using the following formula:
ts = ((X1-X2)/s )*(((N1*N2)/(N1+N2))^0.5)
Where:
X1: corresponds to the mean of the control group.
X2: corresponds to the mean of the experimental group.
s: corresponds to the standard deviation.
N1: is the figure obtained by adding up all the scores of the control group.
N2: is the figure obtained by adding up all the scores of the experimental group
The standard deviation “s” was calculated using the corresponding formula in the Excel program. The range used was a column “d” that contains the results of the subtraction of the scores obtained by the control group minus the scores obtained by the experimental group.
In order to pick up a t-value from the t-chart, a degree of freedom has to be calculated beforehand using the following formula:
DF = N – 1
Where “N” corresponds to the group with the highest numbers of subjects that is being compared with the other group.
Therefore the DF in the case of the male group is 26 and 19 in the case of the female group. A confidence level of 99% was then chosen that corresponds to the followings values in the t-chart (T): T(0.99) = 2,4786 for males and T(0.99) = 2,5395 for females.
The interpretation of the t-score is done according to these criteria:
If ts>T, then there is a significant difference between the groups.
If ts<t, then there is not any significant difference between the groups.
Am I correct in THAT is what the T-Score measure?
That seems right. I would get someone else's approvial as well to make sure, but that looks like you've got it correct.
|
|
|
