My last post has been a while since I have been in San Jose, Oakland, San Diego and Fort Totten this month. That is just the beginning of what our Chief Marketing Officer has referred to as the 7 Generation Games World Tour.

Perhaps this would be better suited on our 7 Generation Games blog but I started writing it here so damn it, I’m finishing it here!

20130826-193748.jpg

In addition to work for The Julia Group, like the site visit I just finished in North Dakota, I’ve been working feverishly on starting up the next phase of game development.

We received a $450,000 grant from the U.S. Department Agriculture Small Business Innovation Research competition for rural development. As you can see from the photo above, the areas where we are working are pretty darn rural. The photo below is the view across the street from the office in Fort Totten. The dirt you see there is the parking lot. They’ve been promising me for the last three years that it’s going to get paved. With the new, proactive tribal council, this time it looks like it might really happen.

nothing but trees

Two things related to this I’m particularly excited about because they pull together both work for The Julia Group and 7 Generation Games ….

The very first Annual Tribal Disability Conference will be held on the Turtle Mountain Reservation on October 3rd with Dr. Longie and I as the keynote speakers. We’re giving a talk with the title, “On the Internet, no one knows you’re disabled”, discussing the great optimism people had twenty years ago about the Internet breaking down barriers and how this hasn’t completely panned out. There will be several speakers talking about their personal experiences with disability – my friend and colleague, Willie Davis, who had a spinal cord injury right before his freshman year of college, an individual with a traumatic brain injury –  and lots of other people that promise to be really interesting.

Western Users of SAS Software conference in November, I will be giving a one hour workshop on Exploring Your Data with SAS Enterprise Guide. We’ll be using the data from our pilot study as an example so this will be the first actual publication of our data. Yes, we did a final report to USDA but that doesn’t really get as much visibility as the WUSS conference and proceedings.

 

Dave Winer set off a firestorm with his blog post asking Why are there so few women programmers.

Lots of people on twitter and his blog called him a lot of unwarranted names, said his straight, white, male privilege was showing, etc. etc.

Just a suggestion — if you find yourself getting all bent out of shape as you read my blog, you might want to read to the end.

Winer must have been in school about the same time as me, and he is right. There were very few women in programming.

In the 1980s,  I was working as an industrial engineer, at General Dynamics, doing programming. Nearly every meeting I sat in, I was the only woman in the room.

It’s not any kind of misogynistic statement to say he worked with few women programmers. I worked with one other woman engineer and no other women programmers.

Fast forward to now. There are certainly more women programmers than there used to be. In the work that I do, statistical programming for The Julia Group and educational gaming for 7 Generation Games, the number of women varies. I see more women at statistics conferences, though most are not in senior positions at their organizations – so I doubt they’d work on a team with someone at Winer’s level.

At the gaming side, when I go to start-up events, I meet very few females who are coding.

Long-winded way of saying there were very few females in tech 30 years ago, and at least in my little spheres, they are still the minority now.

WHY?

disks

Like Winer, I can only guess, but I will make four guesses.

1. Women are not encouraged. I took three required programming classes. My older brother was a computer science major and was always there when I had a question. After him, I can’t think of a single person for the next several years. Partly, I’m here as a fluke. I was nine months pregnant, insisted on still walking through the factory and clambering up on machines to see what was going on and frankly, it was creeping the factory management out so the powers that be sent me to a SAS class to get me out of there.

2. Women feel uncomfortable. Back in the day, users group meetings were great because you got freeware on FLOPPY DISKS. My husband, though, wasn’t too excited about meetings with 60 guys and me. When I attend start-up events now, everyone assumes I am in marketing. It gets annoying to have to establish my credibility in every conversation. Random fact about me, I was the first American to win a world judo championships, so I’m a lot more used to being the only woman in the room than most people. Interestingly, I was at a tournament a few years ago, talking to the women registered in the black belt division. The first woman was a programmer (SQL, I think), the second did maintenance on (unbelievably) legacy code in COBOL, the third was supervisor of a road construction crew, the fourth was a carpenter. Maybe I’m here because I early on got used to being in an all male environment. It certainly seems to be a more common path for women in judo (which is one of the more sexist martial arts in the U.S., far more than mixed martial arts).

3. Some men in tech are complete assholes. Let’s just get it out there. There are men who are just vicious in tearing apart their colleagues. Some are more hateful to women and some are just awful to everyone. I was listening to a podcast the other day where someone said, “If you make (this error) in SQL, you should just cut your fingers off and put them on your desk. Let someone who knows what they are doing write code.” These are the people who stand up in conference rooms and want to point out the slightest departure from perfection in the work done by anyone else, with the insinuation that only a complete moron would have used an ARRAY here.  There are some hateful women, also, but I have met fewer of them. I *think* men are socialized more to suck it up and be tough. Women seem to be bothered more by colleagues who are mean, insulting and abusive than men are, although neither gender finds those people to be a walk in the park. Maybe I’m here because I’m a straight-A bitch when the situation warrants.  If you post nasty comments on my blog, I delete them, think “what an asshole” and move on. As for women who have had their families threatened, I don’t worry about it because no one could possibly be that stupid. Darling Daughter Number Three is the world champion in mixed martial arts . (Her best friend’s picture is below mostly just because we’ve known him since he was 13, we love him dearly and I am so proud of him that he won his second UFC fight this year over the weekend.)

Manny Gamburyan

4. A lot of people in tech lack social grace. This includes me, and there may be a better phrase than social grace, which kind of supports my point. The Invisible Developer is extremely kind-hearted. However, he would say things to the children like, “I can’t believe you are having trouble understand integrals. I figured this out by myself when I was in the eighth grade.” And he would be puzzled when they burst into tears. Often, and this may even be the case with Dave Winer and his blog, they say something completely well-meaning from their point of view, like,

“I wonder why there aren’t more women in programming. Maybe they don’t want to do it because it’s like hunting. You need to sit on your ass and do nothing for long periods.”

and they really don’t understand at all why people get upset. It’s a good bet that people whose strengths are more in the technical realm than diplomacy are more often found in tech fields. Maybe I’m here because I have a very thick skin. Almost every time I get stuck on a technical problem, The Invisible Developer asks, “You want me to do that for you?” I do not give my immediate response, which is, “No, I’m fucking 55 years old. I can figure it out for myself!” Instead, I tell him that if he did it for me, then I wouldn’t have learned how to do it so that sort of defeats the point. (See, I do have some social skills.)

++++++++++++++++++++++++++++++++

If you take away from this that I think I’m better than the average woman and that’s why I’m here, you missed the point. I don’t think those reasons I mentioned are always good. Often, I think I have disappointed my children when they wanted someone to say, “Oh, poor baby” instead of “Suck it up.”

They probably would have liked a mother they didn’t have to caution, “Please don’t say, ‘Fuck’ in front of the admissions people from Harvard.”

My female friends undoubtedly wish I would be more empathetic when their feelings are hurt – and I am very, very fortunate to still have those friends after “not getting it” so many times.

My point is that there ARE fewer women programmers and this is why I think that is so.

Ever since my daughters referred to the day in April when we all take our kids to the office as “Bore your daughters at work day”, it has been clear to me that others do not find me as interesting as I find myself.

My children bored at my office

Based on this profound knowledge, when I was asked to talk for an hour about categorical data analysis, I decided my best plan was to not try to say everything I know about this subject. Hence, I narrowed it down to Part 1 and Part 2, with Part 3 if I have time.

Part 1: Explaining a research project in six pictures.

Part 2: Five lesser-known options 

  1. Fisher’s Exact Test. How to get one. Why you want one.
  2. The difference between the Pearson, Likelihood ratio and Mantel-Haenszel Chi-square
  3. When NOT to compare chi-square values directly
  4. Tests of binomial proportions
  5. Summary tables for multiple variables

Part 3: Three clues that your logistic regression model sucks

  1. Convergence
  2. Model fit statistics
  3. Tests of Global Hypothesis

If you are madly excited to come hear me talk about this, you can come to San Diego on Wednesday, August 21st for the San Diego SAS Users Group or to Oregon Day on October 10th when I’ll be at the Oregon SAS Users Group meeting in Portland.

Or, if you are thinking to yourself,

Ha! I would like to hear about categorical data analysis from AnnMaria for three hours, and no children with mouths full of red jellybeans are going to dissuade me!

Well, then, you are in luck, my friend, because I will be giving a class in Las Vegas, November 12th at the start of the Western Users of SAS Software conference – a class I still cannot believe begins at 8:30 am. I think I said I would do Wednesday because I thought it was in the afternoon.

The only other thing that has gotten me up before 10 am is The Spoiled One’s soccer games – and I love her. So, this will be a rare sighting. I will be the lady with the cup of coffee bigger than my head.

I was planning to finish off with the single most important thing I have learned in 55 years, but life has a way of running away with me. Between meeting with an after-school program that will be using our game next month, a meet-up in San Jose, getting our update 2.10 ready to go, writing a book chapter and three conference papers and scaling up for our new grant – well, blogging just hasn’t been at the top of the list.

nounicorns

(For an explanation of the no unicorns, click here)

…. which brings me to number 54 of the things I have learned

Compartmentalize.

One the questions I get asked most often is,

“How do you get so much done?”

The year I got divorced, I won the world judo championships. The year my second husband passed away I wrote an article in an academic journal, a book chapter and a couple of large, successful grant proposals. I also taught college full time.

It took me a long time to learn that when I was at work, I should just concentrate on work. It also paid off for me in multiple ways. I have a private number only my family has and only for emergencies. As I told my children when they were young,

Unless there is blood dripping on the floor, don’t call me.

No matter how stressful things were at home, I could just concentrate on work when I was at work. While trying to figure out the error in my program, I wasn’t worrying about how this child was doing at school, my husband’s health problems or the argument last night. No matter how much work I had to do, it was relaxing in its own way.

I see mothers in particular who are completely stressed out at work because they are trying to juggle the soccer schedule, order their child’s schoolbooks on line and referee arguments all in between meetings or with a client on hold.

Don’t. When you’re at work, work and when you’re away from work, don’t.

I work in an office downstairs and I feel no guilt shutting my office door and telling my family that I’m busy. The world does NOT revolve around my children and it is not a bad lesson for them to learn that early. I *hate* those Disney movies where the mom misses the winning soccer goal her daughter scores because she is on her cell phone negotiating with a client. I once turned to The Spoiled One during one of those and said,

“She can afford that damn soccer camp BECAUSE of those clients on the phone, you know.”

On the other hand, a lesson it took me longer to learn was that when you are not at work, forget it. I don’t take very many days off but when I do, I read books, go hiking, swim in the ocean and don’t feel bad at all about the grants I’m not writing or phone calls that I’m not taking. I teach judo once or twice a week and when I’m there, I don’t take calls, I don’t check my text messages. It can wait. Every time I come back after taking time off, I’m noticeably more productive. The trick though, is when I have time off to really have it be time off and not just a laptop with a view, which is nice also, but not the same as a vacation.

#55 Children are more worth than they are trouble

At age 55, I have learned that life is seldom black and white. One of the few things I can say for an absolute fact is that in my life, children have always been more worth than they are trouble. If you knew how much trouble some of my children have been, you’d know that is saying a LOT.

I’ve never been one of those stay at home and make play-dough moms, and I have never for half a second regretted it. I’m a lot more of the “Find your shoes, quick, because I have to catch a plane after I drop you off at school” moms.

I never bought for one second that bullshit about how having children forces you to make compromises in your career. I’m typing this in the San Francisco airport on my way home from meetings with staff, an after-school program and Kickstarter backers, and also a few days with Darling Daughter Number One and the grandchildren.

My children are smart, funny, independent, good people. I talk to at least one of them every day and many days I talk to all four of them. If it wasn’t for them, I would not have been to Tunisia, attended plays in Broadway and LA, taught judo in south Los Angeles, gone to the Museum of Modern Art in New York, spent lots of days on the beach, been to Disneyland more times than any human beings needs to go and laughed harder and longer than would ever happen in a board room.

The main lesson I have learned – life is good. Work is good. Children are worth it. You CAN have what you want out of life if you keep trying .

There isn’t an age limit for having a good life. 

 

Not sure what the average number of people to have bailed out of jail is, nor the average number of countries from which people have called you for said bail money, but I’m pretty certain that I have exceeded both of those averages. As I explained to my niece, Samantha, the other day, when people are sitting in jail, they are mentally running down names on a list of everyone they know, like this:

  1. Who do I know that likes me well enough that I can call them at 3 o’clock in the morning?
  2. Which of those people is likely to be able to lay their hands on $500?
  3. Which of them has their shit together enough to figure out how to get it to me? (There used to be a notary public open 24 hours at the international terminal at LAX – there is a reason that I know this. Sadly, I hear that service is no longer offered, victim of funding cutbacks.)

By the time you get to number 3, it’s a pretty short list.

In my last post, I mentioned that I had 14 friends, which is an amazing wealth of friendships. Most of those people would bail me out of jail. Some would laugh about it, some would give me a long lecture on my disappointing behavior and some couldn’t raise the bail money unless they stole it.

That doesn’t mean they are all equal.

My friend, Erich, served in the Marines where he said he witnessed a lot of acts of physical courage. Later, he went on to become college president, school board president and many more honors and accolades. He noted that moral courage in the board room is far less common than physical courage – by the same people.

Twice , I’ve been in meetings where I’ve had people yell and swear at me and be generally disrespectful. Both times it was men who were much larger than me, in an attempt to intimidate me (obviously, they didn’t know me that well.) On one occasion, no one said a word, including a couple of my friends. One later commented,

“I just didn’t know what to say.”

The other said,

I knew you could take care of yourself.

I’ll be honest, it did hurt my feelings and disappoint me. The other occasion, one of my friends banged on the table and said,

“You can’t talk to her like that.”

Here is an interesting fact about that meeting. When he said, that, his two sons, who were also present, jumped up immediately to, literally, have his back. Another friend could not be present, but he sent his two sons with the directions to back me up. They also jumped up. A friend of my daughter’s also stood up and shouted,

“Hey! That’s Ronda’s mom! You can’t talk to her like that!”

In my experience, blood is thicker than water. You hear a lot of people say to their friends, “You’re like family to me.” It’s usually not true.

This is the second part of the hard lesson that I learned. Those people from the first meeting are still my friends. They have helped me out professionally, we’ve had a lot of interesting conversations over a couple of beers and they are not bad people.

#53 There are friends who are like family. Then, there are friends who are like friends, and that’s all right. Just try not to confuse the two.

Contrary to appearances, I’m a fairly private person. You might not think so from some of what I write on this blog, but I challenge anyone to read through here and find personal details like the number of my siblings who have passed away or something hurtful someone might have done to me.

Since I’m almost to 55 – both in terms of my list of things that I have learned, and in age, with my birthday next week, I thought I would break with my usual behavior and give two hard lessons that I have learned.

#52 Don’t mistake colleagues for friends

If you work with someone, you can understandably confuse a colleague with a friend. After all, you spend a lot of time together, you have common interests, you might know each other for many years. You travel to the same events. What more do you want?

In my view, a friend actually cares about you as a person.

Over the years, I have worked at a number of organizations and been on several boards of non-profits. There are people I considered good friends, who I worked well with, accomplished a lot. We often went out to dinner or for drinks together, talked on the phone. When I left that job or organization – I never heard from them again. The first couple of times it happened, I was deeply hurt. Now, I don’t expect anything else. I enjoy the intellectual companionship; I’m gratified by any achievements we have together, and when it is all over, I never expect to see them again and I am not disappointed.

As I sit here, I can think of five people in my thirty-year career that I think of as friends, but I am immensely grateful for each and every one of them. One, I met 28 years ago and I saw her most recently today. One, I met five years ago and, coincidentally, I talked to him today also.

I competed in judo for 14 years and was involved as a coach and board member of various organizations for another 28 years after that. I coached hundreds of students, competed on national and international teams with dozens of teammates, was on boards that served thousands of athletes and coaches. After 42 years in the sport, there are nine people I would call my friends. Coincidentally (or perhaps not), I talked to four of them this week. All of those nine people would bail me out of jail, but only five of them would help me bury the body so that I didn’t get caught in the first place. (That is lesson number two, but you’ll have to wait for my next post for that.)

I am not complaining. I feel amazingly lucky. To have 14 friends in addition to my wonderful family is a blessing. As I said in #7, You ARE your associates – I truly hope the young man who told me we get exactly the friends we deserve is correct, because I have a wealth of wonderful friends.

My point, though, is that 14 is a very small fraction of the number of people I have met in my professional and athletic career. If I had learned earlier the lesson not to confuse colleagues with friends, I would have saved myself some heartache.

Ronda with her confirmation sponsor, my friend Jake

My friend, Jake, as Ronda’s confirmation sponsor

 

Despite what you might think from reading this blog, there is an awful lot about my feelings that I keep to myself, which leads me to

#51 of 55 things I have learned in (almost) 55 years – No one really cares about your feelings all that much.

Maybe your parents do. Perhaps your boyfriend/ girlfriend does. That’s about it. Please don’t write long blog posts full of angst – which the wonderful urban dictionary defines as

 a transcendent emotion in that it combines the unbearable anguish of life with the hopes of overcoming this seemingly impossible situation

Oh, seriously, shut up. Your biggest problem is that you told your toddler to be quiet while you checked your text messages. Now you are going on a 1,500 page self-flagellation of your emotional distress at being a bad mother.

When my children were young and fighting while I was trying to write the next grant that would pay money to put food on the table, I would tell them,

If you don’t be quiet, I’m going to sell you for scientific experiments. I work at a university. I know just the right people to talk to.

I also told them that I would skin them alive and tack their hides beside the door as a warning to their sisters if they didn’t shut the hell up.

Maria at desk studying

More than one person has told me that my children could have believed me and been psychologically scarred for life. That makes me laugh. I think what my children learned was not to waste their time ruminating about what some person might casually say to them.

As for the never-ending stream of posts from people who are seeking “someone who truly understands me” – as I said in #47, It’s Mr. Right, not Mr. Perfect, maybe you are so focused on how YOU feel and what YOU want and what YOU need that you are missing a very good person with whom you could be very happy. Why does someone have to “completely get you” ? I don’t know what the hell that means, anyway. I’m not sure Mrs. Shakespeare “completely got” William, or that Barbara McClintock had a “soul mate” and I don’t think Mother Teresa spent a lot of time worrying if the lepers truly understood her motivation. Freud was wrong. Your internal life is not that interesting. Move on.

 

This is the third and last part of my attempt to explain logistic regression in pictures. You can see a picture of odds ratios here,  and a picture of two charts of predicted probabilities, to compare models, here.

If people only know one chart associated with logistic regression, it is usually the ROC chart, though many of them cannot tell you what ROC stands for (not that it really matters) or how to interpret the chart – which kind of does matter, because it’s useful.

The ROC curve is an abbreviation for receiver operating characteristic curve (I told you it didn’t matter). This is a plot of

SENSITIVITY – the percentage of true positives, the people we predicted would die who did, and

SPECIFICITY – or true negatives, the number of people we said would NOT die, who did not

We actually plot (1 – specificity) by sensitivity. If we predicted no one would die, our rate of true negatives would be 100%. Since we predicted nobody would die, we would be exactly right for all of the people who didn’t die. 1 – 1.0 = 0  so we’d be at 0 on the X axis.

On the other hand, we’d have zero sensitivity. Since we predicted no one would die, we would have zero true positives.

At the other extreme, if we predicted everyone would die, we would have 100% true positives and 0 true negatives. Since 1-0 = 1 , that would be at the upper right corner here.

The straight line is what we would get without any predictor variables, if we just randomly guessed whether a person would live or die. The top left corner, where we have correctly predicted all of our positives and all of our negatives is what we would get in a perfect model.

The more that curve is bowed toward the top left and away from the straight line, the better our model.

Curve with substantial bow toward the upper left corner

Let’s take a look at our actual curve from the Kaiser-Permanente data, where we used gender, age, number of emergency room visits and nursing home residence (yes or no) to predict whether or not a person would die within the next nine years.

From this, we can conclude that while our model is substantially better than random guessing – a conclusion that is consistent with what we saw in our previous charts. We can also see that there is definitely room for improvement. Perhaps future research could improve prediction by including behavioral risk indicators such as amount of alcohol and tobacco usage, as well as socioeconomic status and diagnosis of chronic illness.

So, there you have it  – logistic regression in three blog posts and four pictures.

 

I tweeted that I believed one could explain logistic regression results in three or four charts, and Alberta Soranzo tweeted back,

“Try me.”

Challenge accepted. 

These data are taken from the Kaiser-Permanente study of the Oldest Old,  a sample of 5,986 people who were aged 65-95 when recruited into the study. Participants were followed for nine years, or until they died, at which time it was pointless to continue following them as they weren’t going anywhere.

Last post I showed the predicted probabilities charts for two different models. I pointed out that it was quite clear the first model was superior. Using the same sample of older adults, nursing home status and gender were much better at predicting who died than were race and alcohol consumption in the past year (coded only as if a person drank alcohol or not).

Nursing home status and gender are better than those other variables, but are they actually any good? Are they statistically significant? Is the effect substantial?

Next chart to examine is the odds ratios with 95% confidence limits.

odds ratio charted for each variable

If the odds of living vs dying are equal for people in a nursing home and not in a nursing home, then the odds ratio will be 1.0.  If the odds of people dying are LESS for people who are not in a nursing home (NO vs YES) then the odds ratio will be less than 1.  As you can see, the odds of people not in nursing homes dying are considerably less than those who are in nursing homes. Females have lower odds of dying than males. Being in a nursing home (or not) is a better predictor of dying within the next nine years than is gender.

The dots on the chart are the odds ratio for each variable and the bars extend across the 95% confidence interval. If the bars cross 1.0 then the odds being equal is a value that falls within the 95% confidence limits  – or, in other words, the predictor is not statistically significant.

You can also see from this chart that all four of our predictors are significant. You can also see that people who are older and have more visits to the emergency room are more likely  to die.

One nice thing that SAS Enterprise Guide does is produce a series of graphs when you do a logistic regression. Too many people just skim over the table of Type III effects, say what is significant and isn’t and go on their merry way, which is too bad, because sometimes your graphs are very easy to use to explain your results, even to someone with little technical knowledge.

Take these two examples I came across recently using a sample I was analyzing. I’m interested in predicting who died within a nine-year period, and I have a set of possible predictors including Age at Baseline (the beginning of the nine-year period), gender, race, whether they lived in a nursing home and whether the person drank alcohol.

Look at my first graph, with predicted probability of death at each age from early sixties to late nineties. The lines show the probability of death at each age, holding constant gender and the number of emergency room visits (a measure of health). There are ten curves for the ten combinations of five racial groups by alcohol use (yes or no). For our youngest subjects, the probability of death ranges from a low of around 15% for the lowest group to 24% for the highest. Probability  of death by age with lines for race and alcohol use

[Click on image to expand]

Now let’s look at our second graph, which is, for the same sample, across the same age range, predicted probability, again holding constant the number of emergency room visits but with separate plots by gender and nursing home status. Here there is a very clear difference, with the probability of death for males in a nursing home over 50% even at the youngest age. Females in a nursing home have a predicted probability of death around 27% at the same age. For females of the same age who are not in a nursing home, the probability of death is less than half that.

I like odds ratios as much as the next statistician, but the fact is, the comparison of these two charts is a lot simpler for the average person. Race and whether or not you drink alcohol don’t seem to make much difference in this model. In the second model, gender and nursing home status both have a sizable difference.
nhomegender

[Click on image to expand]

 Your graphs don’t always come out so nicely, but when they do, I recommend you use them instead of trying to explain Type III effects to people who have no idea what you are talking about.

Next Page →