Biology Reference
In-Depth Information
information and look at presence/absence or even degree of osteophyte formation to see if
there is a correlation with age. You might find out that there is a relationship, or you might
falsify your statement by finding out there is not a relationship. On the other hand, a state-
ment such as, “osteophytes on the lumbar vertebrae of males form more interesting patterns
than those on females” is not testable or falsifiable. The determination of how aesthetically
pleasing the formation of the osteophytes is subjective and therefore not scientifically
testable.
Hypotheses should also be the simplest explanation for the observation. Making state-
ments such as “osteophyte formation on lumbar vertebrae is due to the overconsumption
of broccoli during life or the type of mattress slept on” are overly complicated explanations
and additionally, these statements would be difficult if not impossible to test, depending on
the type of antemortem information available. The simplest explanation is also the likeliest
one because it is the least likely explanation to be a coincidence. Even if you were able to
show that people who ate a lot of broccoli have osteophytes (as compared to people who
did not eat a lot of broccoli), it remains a dubious relationship and it does not necessarily
mean that the broccoli is causing the osteophytes. On the other hand, it is much more likely
that based on what we know about skeletal degeneration, osteophytes would result from
increasing age.
When undergoing hypothesis testing, we must define the
independent
and
dependent
variables
. As stated previously, a variable is some characteristic that can be observed or
measured (
Walsh and Ollenburger, 2001
). The dependent variable is the variable you will
be measuring and the independent variable is what will change throughout the study. An
exception would be in regression settings, where you have to measure bone length (the
dependent variable) and real stature (the independent variable), for example. You can also
view the relationship as one of cause and effect: the independent variable as causing a change
in the dependent variable (
Marder, 2011
; but see
Konigsberg et al., 1997
and Konigsberg and
Frankenberg [Chapter 11], this volume). In our example of osteophyte formation, age is the
independent variable and the formation of osteophytes is the dependent variable. The
sample will be selected so that there is variation in the age-at-death structure (each skeleton
you look at will have a different real age, ideally distributed from young adults to older
adults); and you will be recording osteophyte formation. You also believe that age (the inde-
pendent variable) will cause a change to the osteophyte formation (the dependent variable).
Co-variates
In skeletal biology, we often run into situations where the dependent variable (the one you
are measuring) is correlated with other variables. These are known as
co-variates
(
Huitema,
2011
). Typical co-variates with the human skeleton include age, sex, ancestry, activity, diet,
disease expression, traumatic injuries, musculoskeletal markers of stress, etc. Each of these
variables could be tied to another variable inextricably. For example, males in one population
may exhibit more trauma of a certain type (due to interpersonal violence) than females of the
same age in the same population (e.g., see
Smith, 2003
). Biological research with humans (or
any organism) will never eliminate every co-variate, but you need to try your best to limit (or
control for) as many as possible. For example, if you want to just look at age, how do you
eliminate sex from having a confounding effect on the results? The simplest way is to analyze
individuals of each sex separately or use samples that are age and sex-matched.
