A statistician's going rate and needs?


I'm finding myself in the comfortable position of being asked to run stats for people on their articles. I've been paid in three different ways:
  • Authorship
  • Money
  • Combination Money and Authorship

As I do more of this work I have a few questions.

  1. When receiving money for a project what is your rate? Is it per job or hourly (I find some jobs are hard to predict how much time some jobs will take and the client often asks for additional work)? I am US and the dollar is my only frame of reference. Since the US dollar is recognized world wide I would prefer the frame of reference around a dollar (though I can do the conversion if need be).
  2. Do you usually work by verbal agreement or some sort of written agreement?
  3. Do you charge the same for visualizations as for stats work (I'm becoming known for my visualizations around my department)?
  4. What materials do you ask the client for (on my last job I ran stats on variables that were incorrect because I did not have a variable key to decode their meaning)?
  5. What items do you give the client (summary write ups, program scripts, related articles etc.)?

Thank you in advance.

PS Bryan and others looking to make side money, I'm finding this is a way to do this and be able to choose jobs you'd like while declining those you don't have an interest in.


Probably A Mammal
Not that I really have experience in this area, but I do understand some of the environment.

1. The rate depends on the project and time commitments. Some people or assignments may be more aptly approached on a per-hour basis. For instance, tutoring is a good hourly paid job. But you could still turn it into a project paid job if you, say, have them pay for a whole semester's worth of tutoring with a commitment to deliver 3 sessions a week. See the difference? It depends on the assignment. In any case, project pricing is still fundamentally based on an hourly wage you expect to be paid and the time you expect to put in. So if you tutor for $20 an hour. You expect a weekly 3 hourly sessions to run $60. For a 16 week semester that's $960. However, you can give an obvious discount to, say, a $15 or $10 hourly wage because you now have a financial commitment. That is the point behind bulk pay: you get cheaper for more quantity, and as a business entity, you get reliable income over that time frame. (Alternatively would be to give a percentage discount on the hourly fee when bulk purchases of a service are bought.)

2. Verbal is legally sufficient, but that doesn't mean you shouldn't cover your arse. Always have something in print. For instance, you could have a blanket waiver or something that outlines your service and either give it to them (their possession is acknowledgement of awareness) or have them sign it for even more assurance (keep a copy). It really depends on the level of professionalism you're going to have. If a friend is paying me under the table, so to speak, for training them at the gym, I'm probably not going to have them sign a release of liability form! On the other hand, when I was training, it was a mandatory part of my initial meeting to get paper work and assessments out of the way.

3. I don't know if I'd categorize them as different priced products. You're providing a service. They're paying for the service. That service comes with certain deliverables unique to the project. This goes back to 1. If visualizations take more work for you, then you're probably going to charge more for that work (your time is money, literally, and it takes more time to do them, so it costs more--the math don't lie!). With that said, you could always market your visualization techniques and offer it as a special service that costs more. Fundamentally, as I said, it goes back to your time, but the nice thing about visualization is that it's a tangible deliverable that has more 'punch' than, say, delivering a cleaned up data set or a report on some analysis. The distinction is really up to how you want to sell yourself.

4. This should come as obvious to me: whatever you need to do the job! Experience will teach you what that is. As you learned, you need to know your data. That requires that the client knows their data! It also requires that they give you everything you need to know to do your job, but you can't anticipate or even rely on them to know that. It's a learning experience, no doubt, and it depends on the project. The real key is that you can still communicate with your client to get what you need to know. It just comes as more professional if you get everything up front. Your having to set up another meeting to get basic things you require to do the job shows inexperience on your part. Live and learn! I would follow this maxim, though: get as much information as you can. Even if you get information that may seem useless, it's better to have it and not need it than to be suck needing information you didn't receive. You may also find that there were project requirements you did not anticipate, and by getting more information about the project, you're already in a position to handle those other requirements. Overly prepared is being prepared enough!

5. Deliverables depend on the project. I remember tagging along on a consulting project with a statistician in our department. We were looking over the report provided by a company about their product passing our policy standards. In the end, the consultant whipped up a report detailing what this company should expect to cover and where they went wrong along with what the department, as reviewers, should look for in assessing these reports. On the other hand, you may do a job that requires you give concrete information such as a report of an analysis methodology, the results of a model review, or a visualization of the data. You may find yourself needing to clean up someone's data, so your deliverable is the data. This is something that should be determined in your initial meeting along with getting project requirements. So just make sure you spell out clearly what it is you're doing for the client. You don't want to think in abstract terms. You want to have quantifiable results. You need to spell it out exactly. Going back to my personal training experience, I could work with a client and talk about "we're going to help you lose weight." That's great, but what does that mean? We have to set quantifiable goals: "we're going to get you down 15 lbs over the next 3 months." That is a task that can be worked on, that can be assessed throughout the assignment, and can be determined to have been accomplished or not. If you cannot do this, then the project is ill-designed. I think you're good at this. Just look at your learning thread. You started it off spelling out exactly what you wanted to accomplish with intermediate steps detailing what needs to be done along the way and you can specify when you've accomplished them. But notice that you also didn't anticipate the breadth of the project, and you made adjustments along the way. While that is always going to happen, it is the experienced professional that learns to anticipate those changes and has the experience to see the full breadth of the project. Experienced or not, just remember not to let your client know the breadth of your ignorance! Always sell them on you being fully competent. You just have to be creative in dealing with adjustments and maybe getting more information than you require, as I outlined earlier.

That's my 2 cents anyway.

I would like to do some work like this, but I need clients. I'm not at the universities, so that makes it a bit hard. I actually wouldn't mind working for a marketing firm because they'd have interesting data to deal with outside of my experience (because I need more experience!). Pretty much any sociology oriented statistics would be good. There's an institute affiliated with Sac State that does stuff, but they don't have any job openings. They take GSRs (graduate student researchers), but then I'd have to get into the graduate program with sociology (which I'm not entirely against if that could pay for my schooling and stuff). I may contact them about internships, just for the experience. I know for a fact I could help students with their data processing, but they probably don't deal with R at all. I think the psych program uses SPSS or Stata. The econ department uses eviews and possibly SPSS or Stata. I wouldn't be much help in that part of their analysis. I'd probably have better luck at UC Davis, but it's a lot harder to find clients since it's a bit out of my way to get there. I wouldn't mind going there for clients. It's the searching and marketing part that would largely be a waste (of time and gas!).