Automated grading of essays

So, EdX is pushing out an <a href=”http://www.nytimes.com/2013/04/05/science/new-test-for-computers-grading-essays-at-college-level.html?pagewanted=all&_r=1&”>automated essay grading system</a>.  This is interesting. The advantages of reliable and helpful automated grading are obvious: Less (tedious) work for profs, faster feedback to students, and (potentially) less pointless variance in the marking (and feedback).

I have a fair bit of interest in this. I’ve been working with Mark van Harmelen on a system for “smooth” assessment, i.e., to make it easier to grade essays. I’ve mostly stayed away from automation of various sorts because I didn’t think it was feasible and NLP is not my thing. With my PhD student Tahani, I’m working on automatically generating multiple choice questions (MCQs), which is probably as hard as grading essays is (for the instructor).

Plus, I love talking about writing but I had grading. Period. I intensely dislike it.

So, potentially auto-essay grading is a good thing. Essay question are fairly easy to write and if they were equi-easy to grade, hurrah!

Let me express some skepticism, even sight unseen. Obviously, they may have beaten all problems and I’ll be wrong. If so, great. My life becomes easier.

But it’s highly unlikely that this is going to be a very useful tool. Consider grammar checkers. Or even spelling checkers. This is a hugely easier task and afaict they still all suck. It’s not even clear to me that they are useful for non-native speakers. I know they aren’t often used by non-native speakers, at least, not the ones who take my classes, even when I suggest that they do. Since one of the things I tend to include in my rubrics is a point for mechanics, I would love it if I could autocheck that. Thus far, I don’t seem able to.

Now, with a very structured rubric and a lot of tuning and a fallback to human mode, it might be possible to do something. But it’s unclear that it will be a net gain and it’s even less clear that it will be instantaneous. Even more so, instant feedback is only a good thing if it helps.

My colleague, John Latham, is a big proponent of Face to Face marking wherein the student hands in a bit of work and you mark it with them sitting there. There’s quite a lot to be said for that. One thing it doesn’t really do, however, is scale very easily.

Calibrated peer review is a really interesting alternative which has the added advantage of training students how to “grade” (i.e., critique) writing. It does seem to scale. So that seems like a more promising alternative.

Delay

I’m trying to keep the bits up every Monday. But Noruz prep and work got in the way of my idea for today which was still on the halting problem. I’ll try to get it done soon!

The Halting Problem

I would have thought that every CS person would know about the halting problem. But they don’t!

The Halting Problem is a key result established by Turing. Intuitively, any universal model of computation (i.e., for every Turing-equivalent system) is such that it can express computations that we cannot determine (in finite resources) whether it will halt or not.

The proof is quite cool and similar to incompleteness proofs or diagonalization proofs (well, ok, it’s an instance thereof). Understanding this class of arguments 1) it just spectacularly cool; these are deep features of the world! and 2) very useful.

(I like Craig Kaplan’s discussion which is just one that came up early in a search. An interesting challenge would be to write up something that could be understood by a general audience. Hmm. Most of what I write on the Bits is motivational, i.e., I’m trying to convince you that it’s important, not explain the Bit per se. Even there, my inspiration for this one ran out: Why wouldn’t you want to know about about the Halting Problem?

Well, if anyone has an idea of what to do, let me know. But it’s better to post than not to post!)

Fitt’s Law

As all my students know, I’m a big fan of Fitt’s law. Again, Wikipedia has a really nice discussion of it (my first encounter came long before Wikipedia existed, via Bruce Tognazzini; alas I’m not immediately finding the article which I read back in the day, but he’s obsessed with it), and  Kevin Hale’s is also excellent.

Very rougly, Fitt’s law states that the time to acquire a target (by pointing) on a linear path is proportional to the distance and the “width” (along the path) of the target. Big close things are easy; far small things hard hard to hit.

Well, d’uh-uhhhhhhhh. Right? Why is this even interesting?

Well, the qualitative statement doesn’t capture the actual, empirically backed, precise formulation. That alone is worth the price of admission. Getting precision out of obviousness is not easy, but doing so is worth it. Knowledge of the law drives a lot of the way interfaces work (e.g., the Mac’s fixed menu bar was explicitly fixed at the top to make it have effective infinite width).

It’s very universal, too. It applies across a wide range of populations and pointing modalities (cf. Hale for more discussion).

Fleshing it out to 2 (and 3) dimensions is also cool work. Exactly the kind of elaboration one would expect from a good basic piece of work.

Hale links to Yet Another Great Article which explores some consequences of the actual details of Fitt’s law.

So, minimally, I’d expect any computer science to at least recognize the name (“Fitt’s law”) and be able to recall the qualitative formulation. It’s a wonderful jewel of computer science and if you understand nothing else about HCI at least you’d have something profound, pervasive, and beautiful under your belt.

Even if you aren’t going to use it in your own work, it makes an excellent case to study for grasping how to go from an “obvious truth” to an empirically verified law, and why we want the latter. It’s really easy to design and even conduct simple experiments around the phenomenon, and it provides an analytical framework for thinking about a wide variety of issues in interface design. So, it’s a great practice area for people looking to study people.

Read up on Fitt’s law and try Tog’s quiz. It’s a great thing to think through.

Here’s a question of my own: Present a scenario where the acquisition advantage of the Mac Menubar is trumped by other factors. That is, when can having the menubar attached to a window (and thus being of finite length) win and win hard? (Use Fitt’s law!)

The Chomsky Hierarchy

CS Bits

This is first in a series of posts on various computer science topics that strike my fancy.

In spite of being a reasonably successful computer science academic, I’ve little to no formal computer science training. So, my background knowledge base is a bit eclectic. There’s a list of things that I presume is common knowledge amongst computer scientists. (I regularly quiz my students on these.)

One reason I think these topics should be common knowledge is because they are super cool. So they seem worth sharing.

The Chomsky Hierarchy

The Wikipedia article is perfectly fine from a technical point of view and for many bits is reasonable accessible. So, I commend it to you and won’t replicate that content per se.

Some key takeaways:

• The correlative ideas of a formal grammar and formal language.
• You should know what productions, terminals, and nonterminals are. You should know, at least roughly the (E)BNF syntax for grammars.
• You should have a firm grip that a language is a set of (finite)?strings, that a (finite) grammar describes that set, and that we can have alternative grammars for the same languages.
• You should know that languages can be, and typically are, infinite (in spite of being described by a finite grammar and containing only finite strings). You should know why and also which infinite cardinality.
• You should understand the dual nature of grammars (i.e. that they can be used to generate recognizers and generators of languages).
• You should understand the correspondence between a grammar and a class of machines for recognising elements of the language.
• From the last, you should see a relation between the expressivity of a grammar and the complexity of recognising the language.
• You should understand how and why the typical regexp package in programming languages are typical not formalisms for writing regular grammars.
• You should grasp the containment aspect of the hierarchy.

And, of course, you should have a good concrete command of the material. E.g. you should be able to write a context free grammar that is expressible as a regular grammar and one which is not, among other things.

Why know all this

First, and probably most importantly, it’s a core part of computer science and one of its glories. Think Maxwell’s equations in physics. Regular and context free languages are the standard way of describing (most aspects) of programming language syntax and every computer scientist should have a basic grip on programming language theory.

Second, studying it gives you a great foundation for a number of skills, particularly in complexity theory. The kind of math you have to master the hierarchy (to any degree!) is a kind of math that’s pervasive in computer science. (Though its not the only key bit of math.)

Finally (for now), it’s great from a conceptual point of view. The very idea that we have languages described by grammars and recognized by machines is amazing and beautiful. It also very nicely models a more general pattern of trying to capture a phenomenon (e.g., a language) by a precise description (e.g., a grammar) that gives you both theoretical insight (e.g., complexity) and a means of computation (e.g., the corresponding machine). While there are limits to this methodology, it can take you a long way.

Personal notes

My first encounter of the Chomsky hierarchy per se was in a lovely little book I could not afford that I read in a Barnes and Noble in North Carolina in the 1990s. It was by Niklaus Wirth (whom you should know of) and, to my great delight, is available online. While dated, it’s still a great entry point into compilation and all that.

A hint or two about the dated aspects:

• Parsing theory has enormously advanced. To be fair, the book wasn’t meant as a treatise on parsing at the time, but you should be careful about many useful assumptions he makes. For example, scannerless parsers (that is, ones without a separate lexical phase ) have received a fair bit of attention recently.
• Compilation has hugely moved on. To pick just one thing, consider the commonality of JIT compilation (e.g., in the Java virtual machine).

That all being said, read it. The prose is lovely and there’s so much to be said for writing a small monograph.

Blackboard Learn 9.x Fail Encore

Last year, our installation of Blackboard could upload grades from a spreadsheet. So you could grade offline! Which is good, because we want to grade off line, esp. programs. But boo! You couldn’t upload feedback so even though the feedback was sitting in a column ready to go, we had to cut and paste it in. BOO!

But then, in spring, a service pack made it possible up load  (and download!) feedback. WOO! This is good! I can grade offline! I can use my tools! I can analyze stuff!

Except I now figure out that if I have multiple question tests, I can’t upload feedback OR MARKS for individuals questions in the test. Which, for something like, oh, I don’t know, a FINAL EXAM is a big deal.

It’s also going to suck for giving feedback. Lots of cutting and pasting in my future.

Software as a service folks of the world, there is a MINIMUM REQUIREMENT on you: Make sure your users can export and import your data. Easily. Very easily. Make it easy, ok? Use freaking XML if you have to. Just make it easy. From day 1. Until day always. For proper bonus points, make sure that simple things can be done simply. But if not that, just make sure we can do it.

Users of SAAS, demand this. DEMAND IT. If they can do it, you should worry.

On Anxiety (2): Speaking and Other People Speaking

I divide my anxiety about public (and private!) speaking into several aspects:

1. My speaking prep: Before I give a prepared talk, I angst a lot. Preparing well in advance is rarely possible and even if I do manage to start in advance, typically I get part way then stop, then finish very close to the talk. Even if I get a polished version in advance, the risk of my ripping it apart at the last minute is very high. I’ve gotten better at just letting this happen. For example, for my classes, I’ll try to review the slides on Monday, fail quickly. Try again on Tuesday, fail quickly, etc. This isn’t ideal, but at least I’m not burning huge amounts of time on Monday and Tuesday. In 2011, I did manage to get through the entire semester without pulling an alnighter. This is a record.
2. My speaking: At the point of getting up, I’m usually ok (esp. if I’ve been furiously revising…it’s actually soothing!). Depending on how long “idle” I am before the talk, I can get fairly worked up. Once I start talking I’m usually fine: I just give up and let adrenaline carry me.This is what deceives people about my calmness. Once I start talking, I’m very fluent, confident sounding, spontaneous, etc. But often I’m just in a kind of fuguesque state. I don’t lose memory, really, but I do often feel like a different person unto not know what the hell I’ll say next. It feels like a bit like automatic writing. (Note, the dissociationesqueness isn’t constant. Sometimes I’m much more in control, sometimes I’m feeling more in control (i.e., riding the wave), and sometimes I’m aware of being pushed (i.e., carried by the wave but I can roughly see where it’s going), and sometimes I’m just jetsam.
3. Other people speaking, my stuff (e.g., students, colleagues, etc.): So, in addition to just direct fright, I start to worry that I slid in an error, or failed to catch something, or failed to prep enough (thus let the speaker down). Or I’m afraid some problem will occur that messes up the talk (I feel this when Zoe performs as well) and expose the speaker to ridicule (which then is my fault).
4. Other people speaking, not my stuff: Why should this cause me any difficulty at all? Because it can. The mechanism tends to be a feeling of crushing (though unsubstantiated) inferiority. If someone is giving a really good talk I tend to have two streams of emotions: 1) enjoying it a lot (I like a good talk; I like learning stuff) and 2) feeling that I could never do nearly so good, so each good book is another nail in my failure coffin. Obviously, if I have given or have to give a talk during the same session, this can interact a lot with the direct stage fright.

Let me stipulate that all this is not rational. I’m a good speaker (I’ve had people come up after a talk and say to me that I was the best speaker they’ve ever seen and they would go out of their way to see me again; this doesn’t mean I’m to everyone’s taste, natch, but it does mean that I have reasonable evidence that I’m pretty good) and I (and my students/colleagues) do interesting work. I have tons of evidence for these facts. I accept the evidence.

But not all of me is sensitive to evidence. That I accept that there is no basis for a fear doesn’t mean that I don’t experience that fear. Over time, I’ve gotten better at limiting the effects of the negative experiences (e.g., I’m much better at not drawing negative conclusions from the fact of a fear experience and not letting what negative conclusions that do emerge get entrenched into my cognitive system). This is great! I’m much less prone to depression now that I generally don’t take my feeling of failure as evidence of failing or evidence of failing as evidence of total worthlessness.

However, it’s still taxing. The emotions themselves are wearying and my coping strategies add even more overhead. (Not sleeping is one of my more effective strategies. I trust the tiring and tiresome aspects of sleep deprivation are obvious.) There are other physiological effects. For example, recently I’ve been having these coughing fits ending with dry heaving that seem to be closely correlated to an anxiety spike. If I dry heave 2 or 3 times on a 20 minute walk to work and spend a chunk of the rest of the time trying to resist it, then I’m going to be feeling pretty crappy by the time I get in.

On Anxiety (1)

A penance post for my extended commenting over the past few days! I still owe an Ada Lovelace day post (I want to post on Suzanne Langer). But this is a nice segue for the On Anxiety series.

I’ve several annoying long term medical conditions. To pick a simple example, I’ve had a form of arthritis since I was about 30. It’s mostly controlled by medication but, for example, my hands are a mess and degenerating. One of the more annoying ones is my anxiety, particularly my social anxiety.

People find this really hard to believe because, ahem, I come across as a rather confident person. I speak up a lot (a lot of a lot…too much, probably) and I’m willing to ask questions all the time. I’m a pretty relaxed looking speaker and I can put together a talk out of thin air with no prep that sounds like a heavily prepared lecture. My stamina in comment threads and email threads is substantial.

So when I tell people that I’m painfully shy (about some things) and have difficult to overcome social anxiety they often just deny my claim. They can’t believe it. Which is funny and a bit annoying. (I am, indeed, annoyed a lot but all this. It’s a much better attitude to cultivate than sadness.)

My anxiety really makes it hard to write. Writing is a perpetual struggle. I’ve been trying to write some tiny bits of text for a grant proposal and Robert, my poor co-proposer, has been doing everything he can to extract a mere few words. (Useful words. I know exactly how they should go, usually. But putting them down just doesn’t happen.) Many of my usual tricks (e.g., staying up late/all night) aren’t working.

This certainly was a big deal for writing my thesis (hence the X-teen years it took). But it’s also the case for papers, email, exams, comments, blogposts, slides…

I find it really weird to be simultaneously such a voluble, articulate person and yet utterly mute in key contexts.

Missing the Point

Oy. So many other things overdue at the moment (including an Ada Lovelace Day post), but I thought I’d follow up on this thread from FeministPhilosophers, if only because I made a point that someone (Not Bothered) wanted more detail on and the thread has been (quite sensibly, IMHO) closed to that part of the discussion. I think that Not Bothered has, afaict, been pretty reasonably answered, so I’m not really adding more here. Perhaps repetition will help…who knows.

So, I wrote that e.g., Not Bothered didn’t get Rebecca Watson’s original point about the elevator encounter. My understanding of Watson’s fundamental point:

A few days later, I was making a video about the trip and I decided to use that as an example of how not to behave at conferences if you want to make women feel safe and comfortable. After all, it seemed rather obvious to me that if your goal is to get sex or even just companionship, the very worst way to go about attaining that goal is to attend a conference, listen to a woman speak for 12 hours about how uncomfortable she is being sexualized at conferences, wait for her to express a desire to go to sleep, follow her into an isolated space, and then suggest she go back to your hotel room for “coffee,” which, by the way, is available at the hotel bar you just left. [emphasis added]

Note that this is a fairly minimal claim and one that seems pretty obviously true. In this piece, she points out that it is also likely to be highly counterproductive to the man’s goal (assuming that the goal was companionship and creating a hostile environment). So this is, to people of good will,  helpful advice. I would generalize it to “Don’t hit on people at conferences”. That’s not the point of the conference. I also strongly suggest not hitting on your students. Similarly, co-worker romance has pitfalls. One doesn’t have to have strict rules about it, but it’s not hard to recognize that treating venues which are not designed or intended for romantic pursuits as if they were is problematic. It shouldn’t be hard to see that infecting all situations with a dating (to be charitable) ethos can be rather disruptive, esp. to women, esp. since “being available” is something women have to struggle against in all sorts of contexts.

So, did Not Bothered miss this point? Let’s pick the first comment:

Briefly, I’m bothered by the kind of victimization Skepchick’s original response pins on women. It presents them as beleaguered by the evils of patriarchal society. ‘Oh noes, the menfolk are hittin’ on me.’

I’m not sure if this is missing the point or just “just” disagreeing. It certainly doesn’t seem to be very charitable.  Just consider the routine sexism Watson experienced:

But after a few years of blogging, podcasting, and speaking at skeptics’ conferences, I began to get emails from strangers who detailed their sexual fantasies about me. I was occasionally grabbed and groped without consent at events.

I don’t see how this isn’t worth calling out as wrong and I also don’t see how Not Bothered is going to not classify this as whiny victimisation. (Mostly because that seems to be the schtick in play.)

For he’s calling attention to how ludicrous it is to characterize this event as a case of anything other than an awkward social interaction. For Skepchick to feel as though she has some obvious entitlement to apology here strikes me, in this context, as false.

Did she ask for an apology? Not that I saw. (And why isn’t she deserving of an apology. If you come on to someone in an inappropriate way or situation, and apology seems the minimal thing to do.) The point isn’t to apologize to Watson, but that such actions collectively undermine the space and community for women. And really, if you can’t avoid being hit on after talking about the problems that people treating conferences as places to pick up women for hours, when can you avoid being hit on? The extra problems with the interactions (i.e., the cornering, isolation, overriding specific expressions of intent by Watson) are important, but not essential to the problem.

But Skepchick’s original characterization of the event makes it clear that, prior to the rash of ignorant abuses she’s received after her video, the only threats she faced were the ones in her head.

So, the description starts in her video at around 4:32. Here’s a transcript which I’ll excerpt:

 Um, just a word to the wise here, guys, uh, don’t do that. Um, you know (laugh), uh, I don’t really know else to explain how this made me incredibly uncomfortable, but I’ll just sort of lay it out that I was a single woman, you know, in a foreign country at four AM in a hotel elevator with you, just you, and I don’t wanna go back to your hotel room after I finished talking about how it creeps me out and makes me uncomfortable when men sexualize me in that manner. So, yeah, but everybody else seemed to really get it, and thank you for getting it.

Note that there is no explicit discussion of threat per se, though some subsequent commentary (esp. by Amanda Marcotte) emphasises the threat aspect. Even if Waston wasn’t frightened by the incident doesn’t mean that the threat was not there and the situation wasn’t problematic. Indeed, even if the threat was “in her head” that doesn’t mean that the problem was with her. (After all, she wasn’t attacked, so there was no materialised threat. But that doesn’t mean that she didn’t have to deal with it as if it were a threat. And note that the threat needn’t be of being attacked physically, but there’s also the possibility of the person pressing harder, of insulting her, etc. Given her experiences and the events leading up to that moment, it is quite rational for her to have such concerns. Given that she was exhausted as well, makes being extra creeped out even more understandable.)

Not bothered gets much worse:

Instead, she takes an awkward situation and makes it worse by posting an indignant video characterizing herself as a victim and he an aggressor. She had a late night with some friends, and one of them asked if she’d like to continue the conversation over coffee in his room. Heaven forfend! What a travesty of the patriarchy!

Watch the video…it doesn’t read as indignant to me but exasperated. It was exasperating. What does she have to do not to get hit on, explicitly say, “Ok, I’m heading off to bed. I don’t want to be hit on!”

Even as an isolated incident this would be wearing, esp. in the context. Did the guy listen to anything said? What was he doing hanging around if he didn’t understand the issue? However, as the Slate article adequately points out, this was part of a large pattern. And it seems very true: More women would feel more comfortable at such events if “romantic” moves were confined to appropriate times and places.

I won’t go any further into Not Bothered’s comments or even the rest of this one. It’s rather hostile and contemptuous in a way I find unappealing, unhelpful, and unlikely to result in any useful dialogue.

It is odd to have to be discussing the original incident in such detail, as Rebecca Kukla (and others) point out: The abuse afterwards (and, well, before!) are horrific. But I do think Watson had a point with the original issue. Of course, the skeptical community is going to seem even less inviting to women if they tolerate the huge level of abuse she’s received or subtly sanction it by focusing overmuch  on the original incident and compound it with the sort of stuff that Dawkins and Not Bothered push.

The Rape-is-OK-for-the-Survival-of-the-Human-Race Thought Experiment

A rescue from the past, which seems relevant to this discussion.

A surprising (or not so surprising, alas) thought:

So, the premise is that the human race is nuked almost to extinction by a malevalent race of robots called the Cylons that humans created decades prior. …

Then you mix that with the whole human-race-on-the-run and barely surviving, while we try to scrape by and the “Oh yeah, we should make babies so we can propogate the species” thing. Man, it is intense. It makes you think about everything in a different light.

For example, in real life, I’m pro-choice, but if the human race was whittled down to 50,000 people, I think I would want every baby that we could possibly make. For all those closed-minded people out there, it really makes you consider things in new ways.

Battlestar Galactica presents a (vivid) thought experiment. I’d presented a variant to a friend of mine (male and a sort of liberal feminist, but he has read Marilyn Frye) who, like the commenter, tended to cave at the prospect of no more babies (OTOH, I’m pleased to say that my mom was resolute for No Raping To Save The Species).

The line “I think I would want every baby we could possibly make” does not directly call for coercion, though the rest of the paragraph suggests it strongly. Let me rephrase it in a stronger way:

Suppose there were only 50,000 human beings left. In this case, would it be permissible (or required!) to:

1. forbid women from deciding to abort (or not to abort; for example, the policy might dictate that male fetuses were aborted very early or the policy might require extremely low risk tolerance for threats to the mother or the policy might be strongly against letting any by the most perfect, healthy fetuses come to term),
2. compel women to conceive, or
3. compel women to do little else but conceive and carry to term babies.

Here’s another variant of the thought experiment: Suppose that every woman living decided spontaneously and autonomously that they did not want to bear any more children (let’s assume that all current pregnancies that were priorly intended to go to term will go to term, so there’s no abortion component).

1. Is it permissible or required to force (some) women to conceive (and bear) children.
2. It is permissible or required to force women to conceive (and bear) enough children to sustain
   1. current population,
   2. current population of each self-identified identity which desires survival of that (partially birth determined) identity,
   3. a smaller population that was sustainable (i.e., had sufficiently robust genetic diversity), or
   4. a population like that we would have if women hadn’t chosen not to conceive?

Flip the sex: Suppose all men autonomously and spontaneously decided that they did not want to spawn children. It is permissible or required to force (some) men to impregnate women either through intercourse or through artificial insemination?

My answers to the questions in the first two scenarios is “Obviously, no! are you crazy?” though I suspect that it’s stronger in the second scenario. That the first and second case would prompt people to think, “Weeeelll, maybe”, or “Heck yes” is some sort of indicator, I think, of an non-feminist attitude. The first is perhaps less problematic for people given that it’s generally accepted that societies may coerce people into working, killing, or dying in order to ensure the survival of the society. Of course, there are limits there: While existential threats are held to permit more coercion and (intragroup) violence by the defending society, one might hold that there are limits. (If we were actually hit by a annihilating first strike, would a return strike be actually justified? I think not.)

Of course there are other ways to get women to bear children (e.g., economic and social incentives, of varying degrees). (Go mom! She nailed this by the time I’d gotten to the third case.) I would have thought that if one is strongly invested in the continuation of the species, then one would be willing to pony up. Offer huge bounties and lots of social status or political power or the like. In the Battlestar Galactica case, it may be that the marginality of the society precludes some of that (though I’m skeptical). If you jump to the coercion without checking out whether other incentives are possible, that’s an indicator of something worrisome, in my book.

The third case does have a point of disanalogy: the relatively lessor burden (no child bearing). Obviously, the rape variants are off (forcing intercourse, or forcing masturbation…imagine that the men stopped for religious reasons, or imagine that they contracted a disease that meant they could only fertility ejaculate when tortured horrible; no way that’s permissible). However, it might be ok to steal their sperm. If they freely ejaculate it might be ok to collect the result, even against their will or desire. As mom put it, we can use all sorts of things we find in the trash.

A metapoint: This is a horrible thought experiment. I put the brutal summary in the title to be clear what is being entertained. I suspect that there are plenty of “tough minded” people, and even not so “tough minded” folks who will be inclined toward thinking “loftily” of “saving the human race” blah blah. I’m hard pressed to think of those folks as feminist, or pro-feminist. The feminists I admire and strive to emulate would be, I think, tough minded in favor of women.