Affirmative Side Bias
First, I want to congratulate everyone for a fantastic season of NFA-LD in Southern California. The activity is growing by leaps and bounds and there are some phenomenally talented debaters in our area.
I’m going to use this forum to discuss one of the biggest problems that NFA-LD had this year – affirmative side bias.
What is side bias? Simply put, side bias consists of factors which are external to the arguments articulated in a debate that give a debater an artificial advantage based solely on what side of the resolution they are advocating. A good analogy can be found in NFL football. When two teams go to overtime, they use a sudden death format – first to score wins. Hence, there is side bias in that the team that wins the coin flip to determine who gets the ball usually elects to receive the kick, and thus has the first opportunity to score. The rules are biased in that team’s favor.
Side bias is reflected in statistical trends that indicate that one side consistently has a materially higher winning percentage than the other throughout the course of the season. In other words, affirmative side bias simply means that the affirmative wins more often than the negative. Negative side bias means that the negative means more often than the affirmative.
The best resolutions have no side bias whatsoever. They favor neither the affirmative nor the negative, and over the course of an entire season, the win percentage for each side rounds out to about 50%, plus or minus a percentage point or two. However, rarely is a resolution that straightforward; most have at least a degree of side bias. With that in mind, the goal of any topic committee should be to minimize the amount of side bias inherent in a given resolution.
Why is side bias bad? The answer is simple – it artificially skews debates. If two debaters are fairly equal in terms of skill, the round may very well be determined by whichever debater has the side skew going for them. In tournaments with an odd number of preliminary rounds, side skew can be particularly unfair. And of course, side skew in elimination rounds can lead to “coin flip decisions,” where much like in the NFL, the person who wins the coin flip has a better chance of winning.
This season, I perceived that the topic would have an affirmative side bias. I’ll discuss the reasons for this in a moment. In order to test this hypothesis, I decided to go back through all of the tournament results and go round-by-round to see how many debates the aff won versus how many debates the neg won.
For the purposes of this study, I compiled results from eight tournaments in Southern California: LD in the IE at Riverside, APU Cougar, Free LD at IVC, Fall Champs, Sunset Cliffs, Socal LD Champs, Cowboy Swing, and Spring Champs. (I did not count the Watson-Lancer as the entry size was too small). For each tournament, I counted up the total number of aggregate aff and neg wins for all preliminary rounds and all elimination rounds. I then further subdivided this total into open division rounds and combined JV/novice division rounds. Then, I simply worked out percentages to yield the results below.
Please note that I probably made errors, although none should be substantial enough to skew the results (although there are an odd number of ballots when there should be an even number). I also excluded byes from consideration as no debate actually occurred.
DATA FOR PRELIMINARY ROUNDS
Aggregate aff wins: 324
Open aff wins: 183
JV/Novice aff wins: 141
Aggregate neg wins: 241
Open neg wins: 148
JV/Novice neg wins: 93
Aggregate aff win percentage – 57.3%
Open aff win percentage: 55.3%
JV/Novice aff win percentage: 60.2%
Aggregate neg win percentage: 42.7%
Open neg win percentage: 44.7%
JV/Novice neg win percentage: 39.8%
DATA FOR ELIMINATION ROUNDS
Aggregate aff wins wins: 47
Open aff wins: 26
JV/Novice aff wins: 21
Aggregate neg wins: 29
Open neg wins: 20
JV/Novice neg wins: 9
Aggregate aff win percentage: 61.8%
Open aff win percentage: 56.5%
JV/Novice aff win percentage: 70.0%
Aggregate neg win percentage: 38.8%
Open neg win percentage: 43.5%
JV/Novice neg win percentage: 30.0%
DATA FOR ELIMINATION ROUNDS, EXCLUDING SUNSET CLIFFS*
Aggregate aff wins excluding Sunset Cliffs: 43
Aggregate neg wins excluding Sunset Cliffs: 16
Aggregate aff win percentage excluding Sunset Cliffs: 72.8%
Aggregate neg win percentage excluding Sunset Cliffs: 27.2%
* There was one extremely large statistical anomaly that popped up in my study – the elimination rounds at Sunset Cliffs. For some reason, the aff-neg win percentage was completely lopsided – there were a total of 17 elimination rounds with only four affirmative wins. One possible explanation may be that in open, the negative team was the higher seed in eight out of the eleven debates the neg won. Combined with the usual randomness and variations that are apt to occur, it does help to account for the blip somewhat. To offset the anomaly, I posted separate elim round percentages that excludes the Sunset Cliffs tournament (and provides a much more telling story, in my opinion).
RELEVANT SUMMARY OF THE DATA
As an aggregate, the affirmative won 57.3% of preliminary LD rounds, compared to the negative winning 42.7%. The differential between these is 14.6% – a very large differential given the size of the sample.
As an aggregate, and when excluding Sunset Cliffs, the affirmative won 72.8% of elimination rounds, compared to the negative winning 27.2%. The differential between these is 45.6% – a shockingly large differential given the size of the sample.
CONCLUSIONS FROM THE DATA
1. There was a heavy affirmative side-bias on this year’s resolution. Debaters who were affirmative in any given round were substantially more likely to win than debaters who were negative in any given round.
2. In elimination round debates, the affirmative team almost always won. Now, the naysayers will tell us that each round is different, and that statistics don’t take into account what goes on during rounds. But when you have a gap that is simply gargantuan in size, there can be no other explanation other than that there was a massive advantage inherent to debating as the affirmative.
Short and sweet – in elimination round debates, the outcome was largely determined by virtue of who won the coin flip. I participated in four elimination round debates this year – in every one of them the aff was the team that had won the coin flip and went on to win the debate. At all three tournaments where I qualified to outrounds, I dropped out with rounds that I lost on the neg (ironically, all three were on 2-1 decisions – incredibly frustrating!)
3. The best debaters almost never lose on the affirmative. From personal experience, I lost two affirmative rounds all year long. When running my hydrogen refueling stations affirmative, I never lost. Period. That’s not a typo – I was literally undefeated with this case. Why is this? The best debaters are always ready to defend every aspect of their case. They are significantly more prepared for negative arguments and are rarely caught off-guard by unique lines of argumentation.
4. Conversely, the least experienced debaters are more likely to lose on the negative. In junior varsity and novice divisions, the negative tended to lose much more often than their counterparts in the open division.
REASONS FOR AFF SIDE BIAS
So, now we understand that there was a very real aff side-bias. The aff won way, way more rounds than the negative did. The question is – why? What were those factors that influenced debates so heavily? I would speculate that there were four major influences this year that helped determine the outcome of debate rounds.
1. Breadth of the resolution. To be frank, this year’s resolution was the broadest one I have ever debated. The number of potential cases was huge – there are literally thousands of possible ways to reform transportation infrastructure. As a result, it made doing specific negative research for cases almost impossible without prior knowledge of what was being run. Which leads us to our next point…
2. Disclosure norms (or lack of thereof). With a few exceptions, Southern California was way behind the curve when it came to disclosing cases on the NFA-LD Caselist Wiki compared to the rest of the country. It was a bit of a hassle to find out what people were running, and I had to rely a lot on word-of-mouth. This made negative research harder.
Now, the exceptions. Special mention to Mat Swanson of CSULA, Shantal Voorwinden of Cerritos, David “Bear” Saulet Bey of El Camino, and the folks at Point Loma for putting up full citations for their affirmatives. Your efforts to make LD debate better and more educational are appreciated.
If you include my own wiki page, that meant only five schools participated, however. I hope that more debaters will be inclined to follow their lead and participate in the wiki in future years.
3. Lack of negative ground. One of the more striking aspects of this topic was the sheer lack of generic negative ground. There seemed to be very few viable disadvantages aside from politics, federalism, and economy/spending/inflation, and the counterplans being run weren’t exactly cutting edge. Generics are typically the backbone of any negative strategy, but without them, the neg gets screwed. I would also say that kritiks were not widely embraced by the community. I ran kritiks usually once or twice a tournament, but many judges weren’t particularly fond of them or (in some cases) didn’t understand the arguments well enough to properly evaluate them. Kritiks had the potential to become the best negative strategy on this topic, but it didn’t work out.
A lack of case research combined with very little negative ground probably made for very poor debates. If the aff has some weird case you’ve not prepared for, and your only generic strategy is a spending disad, you’re probably in trouble.
4. Structural reasons – the affirmative speaks first and last, has infinite prep time, gets an extra speech, and has two more minutes of speech time than the negative.
The first of these is pretty obvious, I think – the advantages inherent in giving the first and the last speeches have been well-documented in the debate community.
The second is infinite prep time. While of course, there is no such thing as “infinite” time to prepare, the idea is that you can spend much more time perfecting your affirmative than your opponent can spend time writing negative strategies for your case. A crude example – let’s say you have a tournament featuring eleven debaters from eleven different schools each running different cases, and everyone is allowed only 20 total hours to prepare. If everyone does 10 hours of work on their affirmative, that leaves 10 hours of negative work that must be split between 10 different cases – translating to 1 hour per negative case. So, in each round, there is a prep skew, because the affirmative will always have spent way more time working on the aff strat than the negative will have spent time working on the neg strat. Pretty simple.
The final reasons are that the aff gets an extra speech and gets extra speaking time. In most policy debate formats, each side has an equal number of speeches – usually four, with the negative getting a “block” of two consecutive speeches before the affirmative’s first rebuttal. However, in LD, the negative only gets two speeches and loses the benefit of the negative block. Additionally, the affirmative simply has more speech time – two extra minutes, to be exact. This is one of my major peeves about LD – it is literally the only competitive debate format in existence where one side is given more speech time than the other.
All four of these lend to a debate format that automatically favors the affirmative, regardless of the actual resolution.
CLOSING THOUGHTS
1. While the topic was interesting, the resolution itself left much to be desired. For the second year in a row, it appears that the framers did not take the time to adequately consider the sheer breadth of the topic and the overall division of ground. Amusingly, last year’s topic was so overlimited that debates became stale after the second tournament, but this year the reverse occurred. For me, at least, the huge scope of the resolution was obvious from Day 1. Not only was it bidirectional (due to the use of the word “reform” instead of “increase”), the term “transportation infrastructure” is so broad, and number of potential reforms so large, that it simply exploded the research burden. Furthermore, the lack of negative ground should have been much more obvious – they could have held an hourlong brainstorming session on potential generic negative ground and they probably would not have come up with much more than what was actually run this year.
2. Judges should be more proactive in considering side bias. In high school policy and NDT-CEDA, one of the things that a lot of judges do is they take into account whether a topic has been aff or neg biased and adjust their in-round practices accordingly. If a topic has been particularly aff-biased, they may compensate by “leaning” slightly negative on procedurals and theory, giving arguments related to ground, limits, and fairness more consideration. Additionally, they may decide to grant more weight to disads with a generic link story or be more willing to entertain arguments that they ordinarily would not. There are many ways in which judges can adjust for side bias, and I would encourage critics to do so in the future.
3. Please encourage next year’s topic committee to consider the breadth of the topic and the division of ground when deciding on a resolution. We already know that the topic area is Mental Health Services, and the topic committee has already put out a request for feedback. Learn how to contact them here: http://www.debatecooperative.net/forums/showpost.php?p=12449&postcount=1
This was a long post, so I’m going to wrap it up there. I hope this was sufficiently informative, and I look forward to your feedback.
- Nick
Did you enjoy this post? Why not leave a comment below and continue the conversation, or subscribe to my feed and get articles like this delivered automatically to your feed reader.
Comments
I think that there is a regional differentiation that needs to be taken into account; which is not exactly what Will said, but where that discussion will eventually lead. For Gorlok and Loma there is an influx of judges from regions outside of SO-Cal’s normal operating procedure. It would be interesting to get a break down from people who have cumes of their tournaments to post breakdowns of the splits (like we do in parli some of the time). This information would be vital to the writing of future resolutions, I believe that we can agree on that.
The one exception may be APU because we got some USAFA and UoP love at that tournament so the regional differences were probably blended, but there was a majority of So-Cal influence compared to Loma/Gorlok.
That being said, “THIS TOPIC IS AFF SKEWED”!!!!!!!!!
I didn’t loose an Aff this year (all the affs are posted on the wiki): I ran Space 3 rounds, Natives about 10 times (the aff lost 2 times this year but not with me running it), and I even ran De-Dev and won 2 rounds with it. My downfall was the Neg in the out round I lost, Tariq had a legit aff. Also, I was to lazy to run anything but Kap at Spring Champs… apparently judges don’t like kap in So-cal, but are not willing to vote on Socialism Specification (lawl @ soc spec).
This reflects the problem of judges voting for the K in LD. I think I lost all the rounds I ran kap because the judge either didn’t like k’s (even though they said they were ok with any argument), thought that USFG action isn’t capitalistic, or that I dropped the ‘golden argument’ they agree with and would have voted on even if I had killed it. The K is the last bastion of hope for a Neg with an Aff skewed topic, but Negs don’t get that (at least in so-cal)! Though, I found out that because I ran the K b4 the T that the T doesn’t matter; which is an amazing RFD when the aff was actually not topical (the reason I went 2-2 at spring champs)!
On Aff Biased topics, what am I supposed to do? Run 8 Ts and 3 specs because no one discloses/posts? This strat has a name, it is called “LAME” discussions I have had this past year point to “Neg is the place to be” because of this strategy. I will repeat, that is LAME. I would still vote for it as a judge (because the flow is golden and intervention is bad blah blah blah), but probably knock some speaker points and drone for 4 minutes of the RFD about how much I think that strat is LAME. If neg teams can only win because of procedurals on a topic there must be abuse somewhere, which means the topic is either A) Skewed – which forces judges to equalize by voting for the procedural, or B) the topic writing sucks. I believe that both of these were true for this year and fear the worst for next year.
In the end, procedurals are LAME and Aff skewed/poorly written topics are also LAME, but to a lesser extent then procedurals. Like any good LD debater, and fan of Brian Boitano, I have a plan. Write better topics! Because that’s what Brian Boitano’d do!
This is not a study that discernible results can be drawn from. I would strongly discourage any judge from changing their judging preferences because of the information given here.
See, in debate, typically arguments have these things called a “claim” and a “warrant.” It’s fine if you want to contest my conclusions, and if you have an alternative explanation for these results I’d be happy to hear it. But for god’s sake, don’t resort to using warrantless assertions to dismiss a couple hours of my work out of hand.
Your conclusions are not based upon a reliable predictive model. There are several factors that could also explain the variation in wins beyond side, that your “model” doesn’t account for. In my opinion you need to use OLS to even begin to make recommendations to judges based upon side bias.
re: warrants,
the time skew is actually very much in the negative’s favor. it is very difficult for any affirmative debater to handle 7 minutes of spread-out off case positions in 6 minutes. negative collapses are almost always very simple, and the 3 minutes that the 2ar gets is simple not enough time to recover.
the topic isn’t nearly as broad as you think. nearly every aff on the topic is vulnerable to t-transportation infrastructure. and even now, in deep outrounds against very good debaters, i find myself going for that simple argument in the camp file. because the USFG does not actually reform domestic transportation infrastructure.
my explanation for your results has to do with the nature of LD in socal. first, the judging pool is very anti-spread, more communicative. this is favorable for the affirmative in a few ways – first, the negative is forced to slow down, making the time skew less relevant, and second, judges are less likely to vote on arguments that are won on the flow but dont’ appeal to the judge’s “common sense.” (politics, t, etc.) finally, i think debaters out here also just do a lot less work on the negative – people aren’t cutting politics d/as or really investing in their procedural arguments and are just going for “this case doesn’t solve, judge,” which is cool but ultimately non-compelling.
none of these things applied at loma, and none of them apply at the biggest midwest tournaments. you had a slew of technically savvy debaters, flow critics, and solid generic negative positions at loma. and, voila, suddenly the negative wins 3/4 of the rounds.
fwiw, i debated in 6 outrounds in the course of gorlok/loma. i went 5-1. all five of my wins were on the neg, all of which were against pretty decent debaters (backhaus, hilbert, jessee, hogan, and spiker,) and my one loss was on the aff against furgie. i’m not sure i’ve lost an outround this year on the negative. it’s THAT favorable.
- Will
^and here the negative from UOP wins, imagine u had one paragraph only to respond to this nick, get it?
First, the disclaimer. I’ve never taken a stats class in my life (Taking one next quarter though – hah!). So, yeah, is my “study” really the epitome of scientific quality? Not particularly. I simply noted an interesting pattern and decided to write about it.
Second, I probably should have been more clear in saying that this is applicable only to LD in Southern California and is targeted towards an audience in Southern California. (This is reflected by the name of the website). So yes, if you debate in other regions with different debating styles we would expect to see different results.
Third, Will, you’re correct in pointing out the regional differences. The judging pools vary substantially. (Much to my chagrin – if these rumors of legit flow judges are true it makes me envy you). The structural advantages I refer to all flow aff in front of comm-oriented critics. In front of legit flow critics who like generics and actually understand how to evaluate arguments (gasp), aff side bias can be largely mitigated.
Fourth, is the neg time skew thing a round-killer? If the aff is a slow debater, yeah. But if the aff is fast, efficient, and has the experience and foresight to understand how much time needs to be allocated in the 1AR, it’s definitely doable. I haven’t done speed drills lately given my judging pool, but I was reasonably fast on the high school circuit when I had tournaments going on every other weekend. And trust me, “reasonably fast” in high school equates to blazing speeds in LD for self-explanatory reasons.
Will, I find it absurd that you post such good results and are presumably rather intelligent, yet you miss the crucial point that the negative literally has less speaking time than the affirmative. Regardless of the structure (which immediately favors the aff anyway, giving them the first and last speech, with the negative unable to balance out this persuasive advantage, lacking the neg block that exists in CX), having less time to make your points always puts the negative at a strategic disadvantage.
Aff bias exists in pretty much every form of debate (inexperienced judges tendentiously prefer to “do something” over the status quo), but LD takes it to the limit. Echoing Nick’s sentiment, I have lost just one affirmative round this year, while my neg loss hovers around 60-40.
Another big problem is depth. The negative only has one constructive and one rebuttal. This makes it really difficult to explain nuances of complicated arguments, which pushes debaters toward more shallow, more “winnable” arguments.
Better topics are always good, but LD needs to change structurally. I propose the following as an alternative:
1AC: 6 Minutes
ACX: 3 Minutes
1NC: 7 Minutes
NCX: 3 Minutes
2AC: 5 Minutes
2NC: 5 Minutes
1AR: 3 Minutes
1NR: 5 Minutes
2AR: 3 Minutes
im going to have to agree with will on this. predominately rounds at higher caliber tournaments like Webster and to a degree OSU/otterbein with a flow panel the negative has a higher win percentage. on the issue of time, its not how much time you have its when you have it. the aff is stuck with a measly 3min speech to round out 6 min of neg smack down. a negative that has properly collapsed should on paper come out with the win on the flow if coverage of the last speech is executed properly. though agree more rebuttals would be nice overall, would allow better development of abuse stories, kritical debates and what not. end of the day the side bias comes down to what type of judge you have in the round. if its lay, its aff. if its flow i would say negitive.
So the topic is aff biased, the time skew is neg biased. I would still rather be aff with the affs I have run this year. Namely, teams have collapsed to procedurals that don’t go well for them because I have written my shells out for 1ar/2ar answers on all violations I could find. Even in fast rounds it hasn’t made a difference for me. I have had fast rounds and slow rounds and rounds in between, but no matter the difference in speed the affs always worked out the same. Maybe I should go more for T/E-spec in the 1nr, but I am not a fan of going all in on potential abuse.
In the end, a new challenger (question) appears! Does a skewed topic make up for a skewed time distribution?
“Will, I find it absurd that you post such good results and are presumably rather intelligent, yet you miss the crucial point that the negative literally has less speaking time than the affirmative.”
check your premises. john put it best:
“its not how much time you have its when you have it.”
Nick: SO SORRY I didn’t catch this when it first came out. Absolutely SUPERB work! PLEASE investigate getting this “published” in the formal academic sense.
Sorry, the comment form is closed at this time.
nick –
i’d pay attention to those sunset cliffs outrounds. that, along with gorlok, was easily the most competitive tournament of the year. and i bet if you got a hold of the gorlok splits, you’d find that the neg was equally dominant.
in highly competitive rounds, the neg is the place to be.