Page 1 of 3 123 LastLast
Results 1 to 10 of 21

Thread: Stats and Graphing for the Mathematically disinclined

  1. #1
    Join Date
    Aug 2008
    Location
    Kansas
    Beans
    77

    Stats and Graphing for the Mathematically disinclined

    I was wondering what the best programs to use for those social, behavioral, and heath researchers who have a weak math and computer science background.

    I am a PhD student in audiology and am using OpenOffice as my word processor with Bibus as my referencer. This combo seems to be workin good so far.

    For stats, I have installed OpenStat through Wine which seems like a great alternative to SPSS. PSPP has a nice interface but is far too limited. The R GUI, RKWard, looks nice but it just doesnt seem quite as intuitive for persons with an SPSS background.

    For graphing, I have been using OpenOffice. I may try Gnumeric but I'm guessing its graphs will look about the same. I liked RLPlot but it is kinda buggy. I havent had time to fully explore Grace yet.

    I realize that programs like R and GnuPlot are superior for those that are proficient with them but I am more interested in hearing about experiences with programs that have nice, intuitive GUIs.

    Any thoughts or comments?

  2. #2
    Join Date
    Jan 2008
    Beans
    23

    Re: Stats and Graphing for the Mathematically disinclined

    It is really hard to make graphs with a gui (you pretty much have to enter programming input of some sort) unless your using a spreadsheet program in my opinion, gnuplot is really easy to learn (the basics) I know windows has a gui for it but its basically the same as the command line you get in linux.

    Your best bets are probably OOo calc or gnumeric, though I don't have experience with many gui software

    I'm surprised you get away with using openoffice writer rather than \LaTeX\

  3. #3
    Join Date
    Jan 2007
    Location
    California
    Beans
    350
    Distro
    Ubuntu 12.10 Quantal Quetzal

    Re: Stats and Graphing for the Mathematically disinclined

    For better or worse there's a tradeoff between control and ease of use.

    A speadsheet program (if it can graph what you want . . . might be a big if) is probably the `easiest' or at least the most gui'ed plot maker.

    On the other end is R or gnuplot where you pretty much have total control but need to learn a little command line.
    Euler_fan
    Ubuntu User # 15369 at the Ubuntu Counter Project
    Woot for GPG

  4. #4
    Join Date
    Aug 2008
    Location
    Kansas
    Beans
    77

    Re: Stats and Graphing for the Mathematically disinclined

    Thanks for the responses so far!

    Anyone have any experience with Grace, RLPlot, or OpenStat? Any comments/complaints about these programs?

  5. #5
    Join Date
    Oct 2005
    Location
    Albany, NY
    Beans
    842
    Distro
    Ubuntu

    Re: Stats and Graphing for the Mathematically disinclined

    Quote Originally Posted by mdshaver View Post
    I was wondering what the best programs to use for those social, behavioral, and heath researchers who have a weak math and computer science background.
    As a research consultant who specializes in social and behavioral research, it always scares the @#$% out of me when I see people looking for stats software for those who have a weak math background. I do _not_ expect social researchers to be closeted computer scientists, but I think it's easy to jump to the other extreme as well. In my areas of expertise (childhood development, child welfare, elder care, etc.) I have read more than one paper where the researcher made some bold claims, but the mathematical methodology made me question the validity of their claims. Unfortunately, these papers were all published in respected peer-reviewed journals.

    Quote Originally Posted by mdshaver View Post
    I am a PhD student in audiology and am using OpenOffice as my word processor with Bibus as my referencer. This combo seems to be workin good so far.

    For stats, I have installed OpenStat through Wine which seems like a great alternative to SPSS. PSPP has a nice interface but is far too limited. The R GUI, RKWard, looks nice but it just doesnt seem quite as intuitive for persons with an SPSS background.
    I've not (yet) used OpenStat, so I won't comment on something I know practically nothing about. Depending on the nature of the statistics you may need to run, it is possible that a spreadsheet program such as OpenOffice.org or Gnumeric may very well be an appropriate tool. MS Excel had some famous problems with certain statistical formulas, but the problems generally affected very very large and very very small numbers, not the kinds of numbers often used by social researchers at least. For most of us, spreadsheets are more than accurate enough.


    Quote Originally Posted by mdshaver View Post
    I realize that programs like R and GnuPlot are superior for those that are proficient with them but I am more interested in hearing about experiences with programs that have nice, intuitive GUIs.

    Any thoughts or comments?
    I try to be software agnostic. I try to use the "best" tool for the job, understanding that "best" may be subjective. If all I need to do is build some cross-tabs with a relatively small data set and present a few tests for independence, I may just use a spreadsheet. Remember, there are two parts to our job. First, we need accurate numbers AND we need to effectively communicate our findings to others. Spreadsheets often suffice for the first requirement and they excel (pun intended) at the latter.

    The knee-jerk suggestions for plotting on GNU/Linux are all very nice (gnuplot, etc.) but are generally unnecessary. Here's a newsflash - you can get nice plots from OOo or Excel. We use them ALL the time and our clients are typically state and county governments. The _real_ disadvantage is in reproducibility and adaptability. To me, Calc is adequately flexible, but it can be slow to use when I need to make lots of graphs. Spreadsheets tie you to a GUI. If I have a data set in a spread sheet and I need a couple of quick charts, I'll probably just use Calc. If I need MANY graphs, I dump Calc and drop down to the command-line to use more powerful tools. These tools aren't necessarily better and for one or two graphs they may very well take longer. But, when I need to produce a couple dozen graphs in a single afternoon, the advantages of a programmable tool are hard to argue with.

    There is another trend that I think will become increasingly important for social scientists. Many social scientists and government (US) systems rely on "flat-file" datasystems. An example of this classical pattern is the AFCARS "database" that tracks the progress of children in foster care. These legacy systems are slowly being replaced with more modern relational database systems, which is a good thing.

    In the realm of social/behavioral research I think there should be more training/education for people earning masters and PhD level degrees with an interest in research. PSPP will, over time, catch up with SPSS. R has more than enough power for anything I'll ever need to do and Postgres/MySQL/SQLite are all great databases. With the exception of R, these all come with a user interface (of some sort).

    I think the solution is more training and education, not new tools.
    Last edited by gunksta; April 13th, 2009 at 02:27 PM.
    Please Insert Funny Statement Here.

  6. #6
    Join Date
    Apr 2008
    Location
    Cork, Ireland
    Beans
    119
    Distro
    Ubuntu 9.04 Jaunty Jackalope

    Re: Stats and Graphing for the Mathematically disinclined

    I personally use latex for all my stuff, papers, posters and presentations, and use jabref to manage my references. I'm studying physics though so need it for equations.

    For graphs I use mathematica but understand you are looking for a free alternative.

    Origin is used in our department a lot for plotting data, but have found qtiplot very similar and a very good free alternative

  7. #7
    Join Date
    Sep 2007
    Beans
    207

    Re: Stats and Graphing for the Mathematically disinclined

    Quote Originally Posted by gunksta View Post
    As a research consultant who specializes in social and behavioral research, it always scares the @#$% out of me when I see people looking for stats software for those who have a weak math background. I do _not_ expect social researchers to be closeted computer scientists, but I think it's easy to jump to the other extreme as well. In my areas of expertise (childhood development, child welfare, elder care, etc.) I have read more than one paper where the researcher made some bold claims, but the mathematical methodology made me question the validity of their claims. Unfortunately, these papers were all published in respected peer-reviewed journals.
    I have to agree. Things like statistics and probability analysis can be confusing and easy to trip up even for people who should know better. You really have to have a good feel for the data before you start an analysis. For example, if you're not reasonably confident that your data consist of statistically independent events, you can really get into trouble throwing the "usual" analysis tools from a software package at it.

    I don't know how common it is, but when I was in grad school, the math department offered free statistical consulting to anyone, particularly for those working on their dissertations. My guess is that it was their way of training their grad students.

    Quote Originally Posted by gunksta View Post
    Here's a newsflash - you can get nice plots from OOo or Excel. We use them ALL the time and our clients are typically state and county governments.
    This is probably a personal preference thing, but in general I would have to disagree, at least for Excel (I very seldom use OOo to make graphs). I find the default Excel graph to be very ugly, and at times, hard to decipher (even when I make them myself). The appeal is that you can make graphs very quickly in Excel, and I will do this myself at times if I want a quick look at data, but it falls way short (in my opinion) for creating publication-quality charts. I find that plot area is too easily sacrificed to title, subtitle, and axis label space whereby the data become secondary to the data labeling. I also have a lot of other issues with it that I won't go into. Most of the issues that I have, I'm sure, can be solved by spending the time to change and tweak a bunch of values (axis ranges, font sizes, symbol types, line types, etc., etc., etc.), but by the time it takes me to screw around with all that, I can have a very nice looking plot done up in R, plus have a very handy script available to regenerate it or make anew with new data.
    Quote Originally Posted by gunksta View Post
    In the realm of social/behavioral research I think there should be more training/education for people earning masters and PhD level degrees with an interest in research.
    Unless my experience with coming up through the physical sciences was unique, I would say the same for the physical sciences as well. We were taught many mathematical techniques for solving theoretical problems, such as boundary value problems, but we were given very little instruction in data analysis. You usually ended up learning it when you started generating your own data and you had to start muddling through it yourself.

    As a funny example, I know of one case where a student was presenting a thesis defense where a graph was shown with two data points that were measured. The next graph was the same data plot with an empirical model overlaid on it that showed very good agreement between model and data. The empirical model that fit the two data points so well was a high order polynomial.

  8. #8
    Join Date
    Aug 2008
    Location
    Kansas
    Beans
    77

    Re: Stats and Graphing for the Mathematically disinclined

    I don't disagree that it is important to have more than a superficial understanding of statistical application before analyzing a dataset; however, this understanding and proficiency with the command line and languages such as R or S do not necessarily go hand in hand. Although I do feel that I am rather weak in this area as I am still a student, I feel comfortable in pointing out that commercial stats and graphing programs which use GUIs are the norm, not the exception. It is apparent that although powerful and free applications such as R and GnuPlot may be a godsend to some it seems highly unlikely that they will be adopted by the masses who may be more comfortable using SPSS and Sigmaplot or similar programs. However in the linux world, there is no need to reiterate how wonderful and powerful these programs are because these opinions are abundant when googling "linux" and "statistical programs". My main interest was in getting feedback from researchers and students who had experience with some of the GUI-based programs (including spreadsheets) and what your thoughts on the advantages and disadvantages are for said programs.

    Thank you for all your responses so far.

  9. #9
    Join Date
    Oct 2005
    Location
    Albany, NY
    Beans
    842
    Distro
    Ubuntu

    Re: Stats and Graphing for the Mathematically disinclined

    The most advanced FOSS statistical package with an integrated GUI is PSPP/PSPPIRE. Rkward is a front-end for R, but it's development may be and writing a GUI for R is a little like putting a tow package on a Ferrari.

    Unfortunately, PSPP is not feature complete. For many of us, it is not adequate. However, I think PSPP is incredibly important because it copies the layout of SPSS so perfectly. As PSPP becomes more feature complete, users with experience using SPSS will increasingly find it to be a competent alternative. Hopefully the next version, rumored to be 0.7, will include an updated output engine and will be a better candidate for showing people who are interested in using Linux but need point and click statistics.
    Please Insert Funny Statement Here.

  10. #10
    Join Date
    Sep 2005
    Location
    Los Baños, Laguna
    Beans
    396
    Distro
    Ubuntu 11.04 Natty Narwhal

    Re: Stats and Graphing for the Mathematically disinclined

    There's always SAS JMP for Linux. It offers a point and click interface that you seem to desire.

    SAS offers a free 30 day trial to see if you like it. I've used it before just for kicks, but stick with R.
    Earlycj5

Page 1 of 3 123 LastLast

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •