Excel Statistics 38: Data Analysis Add-in Rank & Percentile

Hi,

I'm new to the message board and hoping someone can help me out.

I have a list with number of days from a user registered for an exam and when they actually took an exam. I'm trying to figure out what the %'s are. For instance 5% of the time a user registered and took the exam in 10 days, etc. In excel under data analysis I used the Rank and Percentile analysis but not sure that is correct or if there is a better way.

Thanks and if there is more information needed please let me know.

ps ultimately we are using this data for forecasting.

I am looking to run certain statistics in Excel and I was wondering if this was possible:
I want to calculate a certain number of people in a category on a spreadsheet by age and race and I wanted to know if it could be done in Excel. We do not have a specific program here to calculate statistics like SPSS.
I have a spreadsheet of patients with their information such as age, telephone number, Date of birth, race, location of cancer screen, date of cancer screen, and results of cancer screenings.
I have a lot of information and I was interested to know if I could select only certain columns (age, race, cancer result) to run the statistics. I need to make charts containing the number of people who screened for cancer and out of those people, how many were which race and how many were in each age group.
If this kind of data is possible to obtain in Excel, how do I go about doing that?
Thank you!!

Hi,

I have a set of figures for production for the past 18months and I would like to do some sort of forecast and get an idea on what the future holds.

I have tried FORECAST and TREND and also used the TRENDLINE in the charts but the problem is I don't know what one I should be using and which is the best.

Can anyone give any advice or point me to a site that has a "daft but not an idiots guide to data analysis with excel ", long shot but worth a try.

Many Thanks

Mark

Hi,
I'm working on a statistics schoolproject where I have to use spss and excel.
I have a worksheet with 2 variables: gender and internet use.
(lots of rows with cases downwards).
Gender is a dummy-variable (0,1) and internet use is scale variable, counting hours spent each week.

so to the question:
Which data analysis test can I run to find out that a difference between genders concerning internet use is not because of chance alone (assume equal variances)?

I tried to make a statistics test called
t-Test: Two-Sample Assuming Equal Variances

The output I made is included as a file to this post.


I think the procedure is right, but how do I extract the information from this table to help me answer the question?
Is gender difference in relation to internet use because of chance?

(This is a question relating to Excel 2007).

In 'Data Analysis > Descriptive Statistics' there is the option to output a confidence level for the mean (as a %). Does anyone know if the confidence interval is calculated using a t-distribution or does it assume a normal distribution (as with the CONFIDENCE function)?

Thanks in advance
Rob

I recorded a macro to produce descriptive statistics for a column of numbers. I added the analysis.xla to my VBA references. I get the desired results at the end of the recording process, but I do not get the desired results when I run the macro. Nothing happens unless I add a select cell code at the end. That tells me that the macro was activated, but it does not produce any descriptive statistics. After repeated tries at recording my own code, I copied code from discussion boards. They did not work. I can produce essentially identical results by constructing a longer macro that calculates the statistics individually, but I would like to use the data analysis/descriptive statistics route. I would appreciate any guidance.

Is it possible to calculate multiple percentiles values from an array in Excel?

Currently my data is in columns (with a header) and I would like to calculate the 1,5,10,90,95 and 99th percentile for each column and have them displayed as a single output which I can recycle for more analysis.

Trying to figure out a percentile rank using the following data in column F, ranking them from the best being 100% and the worst being 0%, new to excel and having a great deal of difficulty.

TEAMS PR PA DIFFERENTLY TOTAL POINTS POW-R'-ANKINGS Titans 375 234 141 609 23.15% Ravens 385 244 141 629 22.42% Steelers 347 223 124 570 21.75% Giants 427 294 133 721 18.45% Eagles 416 289 127 705 18.01% Patriots 410 309 101 719 14.05% Chargers 439 347 92 786 11.70% Colts 377 298 79 675 11.70% Panthers 414 329 85 743 11.44% Falcons 391 325 66 716 9.22% Saints 463 393 70 856 8.18% Vikings 379 333 46 712 6.46% Jets 405 356 49 761 6.44% Bucs 361 323 38 684 5.56% Packers 419 380 39 799 4.88% Dolphins 345 317 28 662 4.23% Bears 375 350 25 725 3.45% Cardinals 427 426 1 853 0.12% Cowboys 362 365 -3 727 -0.41% Bills 336 342 -6 678 -0.88% Texans 366 394 -28 760 -3.68% Redskins 265 296 -31 561 -5.53% 49ers 339 381 -42 720 -5.83% Broncos 370 448 -78 818 -9.54% Jaguars 302 367 -65 669 -9.72% Seahawks 294 392 -98 686 -14.29% Raiders 263 388 -125 651 -19.20% Browns 232 350 -118 582 -20.27% Chiefs 291 440 -149 731 -20.38% Bengals 204 363 -159 567 -28.04% Lions 268 517 -249 785 -31.72% Rams 232 465 -233 697 -33.43%

I have a collection of data where the lower the value, the better. When I use Quartile/Percentile, it shows higher values as having a higher rank. Any way to reverse the order, without changing the underlying data (ie. multiply all values by -1)?

Also, I have a graph of the data showing the 1st, 2nd (median) and 3rd quartile and then a particular data set (in this case, company results) over a number of years. Any way to automatically add the Percentrank as a data label for the specific company results? For example, the value 200 puts this company at the 65th percentile. I want the data label to show 65, not 200.

I am working with some questionnaire data. Up to this point, I have been working with regression coefficients to determine that a one unit change in A leads to x change in B. What I now like to do is look at percentile rank change - to see when one moves from the 50th percentile to the 60th percentile in A, it leads to x change in B. I have the scale mean and SD and have been trying the NORMDIST and NORMINV functions, but I cannot get it to map on properly based on the z-score and the regression coefficient for the one unit change I previously computed. Is there an Excel function(s) that can map the percentile rank scale (0-100) onto the coefficient values based on the questionnaire scale (1-7)? I hope I explained everything clearly. Thanks for your time!