Notes on Lab report 2. I assigned 20 points apiece to the following
sections: Introduction, Materials and Methods, Results, Data analysis,
and Discussion. For late papers, I calculated the fraction of an 8-hour
work day that the paper was late, multiplied that fraction by 10 and subtracted
the result from the total points to produce the final grade. Between 8
and 24 hours late was considered one whole day late, as I could not be
sure when papers were handed in after the office closed and the building
was empty.

1) Introduction. Most of the problems cropped up here. NOBODY really
laid out the problem in terms anything like Grand’s paper, namely how is
a prehensile tail useful in solving the problems of traveling, foraging,
and feeding in the canopy? To do this well, you would need to: 1) describe
what the problems are (discontinuities between tree crowns, bending of
terminal branches under the weight of heavier monkeys, inability of small
feet or hands to grasp fat branches or trunks); 2) how these problems affect
particularly animals of the size and weight of a capuchin monkey (smaller
animals have less problems with terminal branch bending, heavier ones have
more problem of staying on top of thin branches); 3) hypothesize how the
prehensile tail could be used to ‘solve’ these problems with special reference
to the variables you will be analyzing in the videos. For this last part,
you should write out a series of ** null hypotheses** and

What is a null hypothesis? It is a statement of what you expect to find
if the variables you are analyzing are NOT functionally related to each
other. Typically, they have the form of "variable X is independent of variable
Y", or "variable X is expected to occur with equal probability regardless
of the value of variable Y". The alternative hypothesis is a statement
of what you expect to find, based on theoretical considerations. It is
not enough just to say that you expect there will be an association between
the two – there are many possible results that would make no sense at all.
Any expected patterns should be justified by theoretical principles or
at least common sense – simply stating that "I believe this is true" is
not sufficient. For instance, howler monkeys should have problems of falling
off or breaking thin branches more than medium or large ones, so you would
predict they would use their tail to grab onto supports more if they are
on thin branches than on medium or large ones. Note that the goal of any
scientific analysis is NEVER to PROVE a particular hypothesis or theory
to be true or false. Real scientists can at best make strong probability
statements, but this is not the same as proof.

The Introduction is NOT the place to anticipate your results in detail.
You may tease the reader by giving a general brief statement of the overall
findings (e.g., "In fact, tail use proved to be strongly associated with
some aspects of the substrates used by the monkeys").

2) Materials and Methods. As James pointed out 2 lab sessions ago, the
M&M section needs to provide enough information so that a reader unfamiliar
with the subject could in principle go out and duplicate your results.
You may cite the lab handout to justify the methods (for lack of a better
reference), but still need to describe the methods, criteria, and any problems
you had implementing them. Note that you may use a given set of data, observations,
video clips, or whatever ONCE AND ONLY ONCE in a given analysis – looking
at a video and recording the data several times does NOT give you a bigger
sample size, despite what you may think. ALL conventional statistical tests
require that the data points be INDEPENDENT – clearly repeating data from
the same video would not produce independent data points, because if you
knew the first set of data, you could exactly (or nearly exactly) predict
what the second set would turn out to be.

3) Results. You need to describe the results in words, not just using
statistics (which give probability values, but not a sense of what the
data look like). You should compare the observed results to the hypotheses
you set out in the Introduction. Note (again -- James already told you
this) that if a statistical test does not yield a value large enough to
reject the null hypothesis, you really have NO reason to discuss the result
further. Exceptions may be made on occasion for patterns that are nearly
significantly different from that expected in the null hypothesis AND in
the direction predicted by the alternative hypothesis. Note that it is
ALWAYS incorrect to state that "the data were not significant". Data themselves
are never significant or not significant; it is only the PATTERN of the
data RELATIVE TO the null hypothesis that can be considered significantly
different or not. Also, you MAY NOT just look at the patterns in the data
and claim that they exist or are significant – you HAVE to back up any
such statements with statistics. The reason is that human brains are INCREDIBLY
good pattern-finders; they are SO good, that they will find patterns where
there aren’t any (in the sense that the data were generated by a random
process). Also, note that an hypothesis of no association between X and
Y does NOT mean that X and not-X should occur the SAME amount for each
value, level, or category of Y. For istance, tail use was NOT expected
to occur half the time for each angle category or thickness of substrate.
The overall frequency of tail use is dictated by the animals themselves
– it could have been only 5% of all the observations, or 95%.

4) Data analysis. Generally you all did the Chi-square calculations
correctly, given the format of your tables. However, several did not follow
instructions in handout to combine data in original data tables so that
only one ‘factor’ (orientation, angle, OR thickness) was compared to tail
use (yes vs. no). Using more complex tables (e.g., combining substrate
size and orientation or angle) would be OK, but in practice, the people
who used these did not know how to perform the tests and interpret the
results. Extra points were given for use of graphs to illustrate the tables.
Usually, statistical tests produce conclusions that are or are not true
(the results are, or are not, significant at some stated probability criterion),
rather than "somewhat" true (never mind that philosophers debate at length
about partial truth values).

5) Discussion. The Discussion is NOT a place to simply repeat the conclusions
or inferences from the results. A brief summary of the general outcomes
of the Results is appropriate, to remind the reader. You should then have
related the results obtained back to the general principles or questions
brought up in the Introduction. Few students did this, primarily because
few actually provided such general views in the Introduction. As a lesser
fallback, you could compare the patterns of tail use to that found in other
primate species (as described in Grand’s paper or other references available
in the literature). If the results did not come out as expected, here is
the place to discuss why not. Possibilities include that 1) the hypothesis
was in fact wrong -- if so, find the error in the assumptions or logic
that led to an incorrect prediction; 2) you performed the analysis incorrectly;
3) the data were too few for a significant result to be found (but the
pattern of the results does agree with the alternative hypothesis); 4)
more than one hypothesis may be at work, with contradictory expected patterns
(e.g., some physical factors may make it necessary to use tails more on
thin substrates, and other physical factors may require more tail use on
thick substrates, thus producing little overall pattern). In any case,
describe what you would do to improve upon the study design to obtain a
more convincing result (convincing does not mean significant – even ‘negative’
results [accepting the null hypothesis] can be more or less convincing).

6) General stuff: a) again (James did tell you this), DATA is a plural word: "these data show" NOT "this data shows"; b) the possessive form of "it" is written "its" WITHOUT AN APOSTROPHE – "it’s" is a contracted form of "it is"; c) every English sentence (except in ‘creative’ writing) must possess a subject and verb. Verb forms ending in ‘ing’ are not verbs but gerunds, which usually act as adjectives or adverbs. Thus, "The monkey hanging by its tail." is not a sentence. READ OUT LOUD every sentence you write – if you had done this before handing in the assignment, you could NOT have left in many of the ‘sentences’ that you did. I marked non-sentences with NAS (Not A Sentence).