04.17.06
Posted in grad school at 11:24 am by Brandon
Note: For the uninitiated, SPSS is a statistical analysis program that allows users to crunch statistical data. It has been particularly developed for the social sciences, thus the name Statistical Package for the Social Sciences.
Dear SPSS creators,
You are truly dispicable people.
You know full well that a lowly graduate student needs the service you provide. Sure, there are other statistical options. SAS, R, or other statistical options. But SPSS is the only option with a user friendly interface. And, frankly, you know damn well that I haven’t the time to learn another fucking computer language just to figure out if the mean of scores in Group A is statistically different from that of Group B.
Sure, I could calculate the independent samples t-test or one-way analysis of variance by hand, but you know I don’t want to–it’d be a monumental waste of time. God knows, I’ve got a million other things I’d rather be wasting my time on, like beer or Tetris. And, even if I did calculate those simple statistics by hand, they’re of little practical use. You all know as well as I that a 3 x 2 factorial design requires more than a simple one-way analysis of variance. You have me in your back pocket and you know it. Beyond the ANOVAs, you’re also pretty sure I’d never take on the task of running a regression analysis by hand. Once again, you have me. You have a monopoly on the market of statistical software packages and you seem to love to wag your dominance in my down-trodden face.
You knew damn well when you sold me the “student” version of your software package (for $89.95) just a year ago, that I’d be back, needing better functionality. Seriously, a maximum of 50 variables. I shit studies with more than 50 variables.
Then a year later, you have the gall to charge me another $206.84 (including tax) so that I can get the full version of SPSS at the GradPack price. Of course, that full version expires in 4 years–so you know I’ll be back for more.
I loathe you so much that I almost–just for a minute–wished that I was a qualitative researcher (gasp). I now shudder at the thought of what you’ve made me consider.
Dear SPSS developers, I beg it of you that you would take pity on the graduate student. Please, stop raping us at the bookstore. You know that if we get hooked on you as Graduate Students, we’ll keep coming back as faculty members. We really don’t care if you’re sticking it to our department–charging $600 software licenses and such. But, why mistreat us poor graduate students. You do realize that you force us to make inethical decisions about the sharing of your software, right? Why not just make it cheap for us to obtain legal copies of your statistical packages, and then pass the costs on to departmental software registration?
If you continue to persist in this uncouth behavior I will be forced to learn the “R” statistical programming language and calculate my statistics using their FREE non-commercial software package.
Signed,
Brandon
Tags: Grad School, SPSS
Permalink
Trackback URL »
http://www.badchristian.com/2006/04/17/an-open-letter-to-the-developers-of-spss/trackback/
Posted in grad school at 11:24 am by Brandon
Note: For the uninitiated, SPSS is a statistical analysis program that allows users to crunch statistical data. It has been particularly developed for the social sciences, thus the name Statistical Package for the Social Sciences.
Dear SPSS creators,
You are truly dispicable people.
You know full well that a lowly graduate student needs the service you provide. Sure, there are other statistical options. SAS, R, or other statistical options. But SPSS is the only option with a user friendly interface. And, frankly, you know damn well that I haven’t the time to learn another fucking computer language just to figure out if the mean of scores in Group A is statistically different from that of Group B.
Sure, I could calculate the independent samples t-test or one-way analysis of variance by hand, but you know I don’t want to–it’d be a monumental waste of time. God knows, I’ve got a million other things I’d rather be wasting my time on, like beer or Tetris. And, even if I did calculate those simple statistics by hand, they’re of little practical use. You all know as well as I that a 3 x 2 factorial design requires more than a simple one-way analysis of variance. You have me in your back pocket and you know it. Beyond the ANOVAs, you’re also pretty sure I’d never take on the task of running a regression analysis by hand. Once again, you have me. You have a monopoly on the market of statistical software packages and you seem to love to wag your dominance in my down-trodden face.
You knew damn well when you sold me the “student” version of your software package (for $89.95) just a year ago, that I’d be back, needing better functionality. Seriously, a maximum of 50 variables. I shit studies with more than 50 variables.
Then a year later, you have the gall to charge me another $206.84 (including tax) so that I can get the full version of SPSS at the GradPack price. Of course, that full version expires in 4 years–so you know I’ll be back for more.
I loathe you so much that I almost–just for a minute–wished that I was a qualitative researcher (gasp). I now shudder at the thought of what you’ve made me consider.
Dear SPSS developers, I beg it of you that you would take pity on the graduate student. Please, stop raping us at the bookstore. You know that if we get hooked on you as Graduate Students, we’ll keep coming back as faculty members. We really don’t care if you’re sticking it to our department–charging $600 software licenses and such. But, why mistreat us poor graduate students. You do realize that you force us to make inethical decisions about the sharing of your software, right? Why not just make it cheap for us to obtain legal copies of your statistical packages, and then pass the costs on to departmental software registration?
If you continue to persist in this uncouth behavior I will be forced to learn the “R” statistical programming language and calculate my statistics using their FREE non-commercial software package.
Signed,
Brandon
Tags: Grad School, SPSS
Permalink
Trackback URL »
http://www.badchristian.com/2006/04/17/an-open-letter-to-the-developers-of-spss/trackback/
Jacob said,
April 17, 2006 at 4:28 pm
Two words: rhetorical criticism.
Brandon said,
April 17, 2006 at 4:42 pm
Two words: generalizable results.
;)
Stephen said,
April 17, 2006 at 5:50 pm
Two words:
qualitative analysis and critical theory
Ok, 5 words
Brandon said,
April 17, 2006 at 5:55 pm
I’m actually down with qualitative analysis done well. It’s not my bag, baby, but I think that communication needs mixed methods to truly move forward.
I suppose my problem is that some of what tries to masquerade as a social science isn’t really science at all. Not that there isn’t a real and important place for the humanities, it’s just that humanities shouldn’t claim to be science just like science shouldn’t claim to be the humanities.
The moral of the story is this: lots of qualitative research is done poorly. And, the same goes for quantitative research in communication.
This, unfortunately, sucks.
benjamin said,
April 17, 2006 at 10:29 pm
You know, I’ve spent the last year or so going from one shitty project to another - and they’ve all centered around the calculation of ANOVAs, and the generation of the signifigant box plots and trend charts.
But every day I give thanks that I’m using a real language like perl, and not a bletcherous hack like R.
Shannon said,
April 17, 2006 at 11:01 pm
I used SPSS for analysizing the data for my undergrad senior research thesis, and even for that, it was a bitch. Just annoying as all getup, especially when the class that was supposed to teach us how to use it was confusing more than anything else. (It didn’t help that I constantly felt like my teacher didn’t understand my questions, as she would often answer with a totally unrelated comment.) Thankfully, I at least got to piggyback on the department’s license. It meant that I couldn’t do any statistical analysis on my own computer - woohoo, library computers and e-mailing - but at least I didn’t spend 90 dollars for less than a week’s worth of use.
mike said,
April 18, 2006 at 1:47 pm
My wife (who is taking her third course in statistics right now) often says that the activity of Hell is to do Statistics word problems designed by self-important morons. I will now tell her to add SPSS learning curves to that equation.
Leighton said,
April 18, 2006 at 4:20 pm
Two words (after those two words and this parenthetical aside, with an additional word to follow): software piracy.
Seriously.
Zeke said,
April 18, 2006 at 7:58 pm
I take it Excel is out of the question?
When I was an undergrad, I did regression analysis in an ancient spreadsheet program called First Choice, which is absolutely nothing compared to Excel.
Stephen said,
April 19, 2006 at 10:06 am
Maybe I have just been burnt out by too many crappy sociology presentations that use 20 year old data sets with ambiguously defined terms and disproportianate populations that result in overly generalized conclusions.
For example, yesterday I heard a lecture on the differences between “married” and “cohabiting” couples on the likelihood for domestic abuse.
The study used data from 1987-88 that was originally for 13000 respondents, which the researcher whittled down to 6000 for some reason. Of this smaller set, they had about 5400 married respondents and 600 cohabiting.
Their conclusion was that the married couples were less likely to have domestic violence than the cohabiting couples. However, there was no accounting for the difference in the sample sizes, the difference in time together (marrieds averaged 17 years together, while the cohabs wer only together an average of 3-4 years), or the difference in likelihood for someone in a married relationship of 17 years to report abuse as opposed to someone in a 3 year relationship.
Granted, Brandon would never put together such a shoddy study, but it seems to me that quantitative analysis is as open to accusations of subjectivity as qualitative analysis or ethnographies are, but quan. people lay this claim on being “science”.
dorsey said,
April 19, 2006 at 10:39 am
I have no idea what you’re talking about, but I’m behind you 100%.
Funny, that’s the same reasoning I used to vote for W.
Brandon said,
April 19, 2006 at 11:17 am
I’m with you, Steve. I will grant you that sociology studies can be some of the worst (I took a sociology of the family course last semester). Primarily, as you point out, there’s so many variables and there’s virtually no attempt to control for confounding variables that I find such studies have little meaning.
I’m also with you, Steve, that quantitative analysis is OFTEN done poorly, and is AS open to accusations of being either ungeneralizable or unreplicable (two key components in my definition of “science”.)
The thing about the claim of “science” is that even the bad studies at least have the opportunity to be either confirmed or disconfirmed. They’re claiming to be generalizable, and when they don’t generalize upon replication, we discount them a little.
There’s a bit of lore that runs around our department (and admittedly, we’re brutally quantitative) of a grad student (who’s since gone on to be a department chair at a MAJOR quantitative communication program) who was, one day, at a colloquium of a qualitative scholar who by some accident was invited to speak at our department. This scholar presented a case study about communication amongst 6 kindergarteners. When this intrepid graduate student asked the scholar a pointed question about whether or not we should be able to generalize the scholar’s results, the scholar replied that the intent of this study was not generalization. The grad student promptly packed up his things and walked out of the presentation.
When a faculty member later that afternoon asked the grad student what he was thinking–after all it’s a little rude to simply walk out of a colloquium–the grad student replied, “I wasn’t upset or anything, I just doubted that I would ever run into those 6 kindergarteners.”
I think that’s the beef quantitative scholars often have with qualititative scholars. Admittedly, the beef goes both ways, and I don’t think one method can answer all questions equally well. Frankly, I think we need both areas to advance theory that is descriptive, innovative, and that produces testable questions.
I also don’t think we’re necessarily disagreeing on the topic. I think we come from different schools of thought when it comes to communication research and theory. And, I think, this will always be the case in communication research. Unfortunately, I don’t believe that students who are trained in both quantitative AND qualititative methods end up with real expertise in either method. Thus, graduate education will likely continue to focus on one method over another.
Love this conversation, Steve. Perhaps some of it can crop up in our panel discussion!
Nicole said,
April 19, 2006 at 4:31 pm
Shhh! I don’t want to talk about this. I’m working on the lit review for my dissertation, the methods section and IRB are almost done as well. I’m doing quanitative research, but only because when you’re cramming a 5 year program into 4, the best dissertation is a done dissertation.
I think that a balance between qualitative and quantitative research is needed…but that means brining together experts in both that can dialogue with one another. However, I’m a sociology undergrad who’s getting a psychology grad degree, so all I know is that it all has something to do with my mother.
bad pastor's wife said,
April 20, 2006 at 2:35 pm
Everything comes with a price:
I was in a Ph.D. program in quantitative analysis, loving every minute with SPSS, until I followed my husband to another state to put him through seminary. One year after graduating and being a pastor - he tells me he’s gay. Dang. I knew that I should have chosen SPSS over him.
dudley said,
April 23, 2006 at 1:10 am
Ever try STATA?
Science said,
June 9, 2006 at 8:05 am
1. Regression in Excel
2. Ever tried a real science?
Patrick said,
June 20, 2006 at 12:46 pm
I’d seriously give Stata a look, Brandon. If your university has a GradPlan set up for Stata, you can get a fully functional perpetual license for Intercooled Stata 9 for $149 (that’s what it runs here at UNT). Nothing crippled, nothing expiring.
If Stata isn’t an option, go with R. Yes, it’s got a steep initial learning curve, but besides being free it’s more powerful than anything commercially available (SPSS, SAS, S-PLUS or Stata) for most statistical analyses. The big downside is that it’s only syntax-driven.
Robert said,
October 31, 2006 at 8:53 pm
Hey, if you don’t like it, why don’t you try out my new data crunching programming language at
www.my.opera.com/datahelper
Robert said,
October 31, 2006 at 9:01 pm
And if you don’t like my programming language, try out PSPP.
It’s one of those GNU things (open source SPSS).
From what I read, they make no promises about how soon they will fix their bugs and they are very understaffed, since they depend on volunteers.