Hi
On Sat, 23 Dec 2000, G. Marc Turner wrote:
> now is what SHOULD be included. Here's what I have thus far (in no
> particular order):
>
> Why we use statistics, brief history, etc.
> Scales of Measurement
I would ignore this issue, except for crude distinction between
categorical (i.e., nominal) and numerical data. It is largely
irrelevant to data analysis.
> Populations vs samples
> Central Tendency
> Mean, Median, Mode
> Variablity
> Range, IQR, Semi-IQR, variance, standard deviation, etc.
Focus most on SD and variance. I also like to formally introduce
Sum of Squares (SS) as a "measure" of variability as it is
involved in so many aspects of statistics (SD, var, regression,
ANOVA).
> Probability
> Z- Scores
> correlation
> pearson
> spearman
> pt-biserial
> phi
> part & partial
I would ignore the esoteric correlations, most of which in fact
are just short-hand formulas for the Pearson. Put phi with
chi-square (i.e., strength + significance together).
> simple regression
I tend to teach this hand-in-hand with correlation (e.g., r^2 =
SSreg/SStotal, t_r = t_b1, ...).
> single sample t-test
> independent measures t-test
> dependent measures t-test
Teach dependent t as variant of single-sample. That is, compute
quantity D=Y1-Y2 and test Mu_D = 0.
> 1 factor, ind. meas. ANOVA
> 1 factor, repeated measures ANOVA
> 2 factor, ind. meas. anova
> mixed anovas
This is a lot to include in a half-course. I would probably be
happy to just introduce basic idea of ANOVA.
> chi-square
> nonparametric alternatives
> mann whitney u
> wilcoxon
> kruskal-wallace <<<<< Wallis, I think
I'm not sure I would bother in first-level course, except perhaps
to note that parametric tests do have assumptions and that there
are alternative approaches if the assumptions are violated.
> power
I would probably just do this in a more intuitive way based on
the analysis of the relevant formula (e.g., t, F). For example,
t more likely to be significant if: SD small, n large, and Y1-Y2
large. Translate these into operational procedures (e.g.,
standardized testing conditions, robust manipulations of whatever
variable differentiates groups, ...).
> confidence intervals
> effect sizes
> graphing
If by this you are referring to preparation of graphs, then this
might be part of a methods course?
> SPSS
> entering data, saving data
> calculating descriptives
> calculating t, anova, chi-square, etc.
> interpreting output from different procedures
Don't forget about the supplementary computer stuff: directories,
saving/retrieving/modifying job files, printing, .... Some of
this is essential, others incidental.
> I'll have them twice a week for an hour and 15 minutes over 14 weeks. I
> already know that I'm going to have to scale things back, but I also feel
> that one reason so little is taught in the course currently is an
> underestimation of the ability of our students. (New university admission
> requirements and requirements to get into the major have raised the quality
> of students slightly over the past few years, of course I'm probably
> overestimating their ability.)
Yes to the latter, especially for the lower-end of the class
distribution, unless you are far more selective than in courses I
have taught. The tricky thing in stats is to avoid the terrible
bimodal distribution of grades that can result pretty easily.
> I'm currently thinking of how I can incorporate the SPSS along with the
> hand calculations. The room I'll be teaching in has 12 machines around the
> perimeter, so I'm thinking I might split the class in half (approx
> enrollment of 30 students). That way I can have half working with SPSS
> while the other half calculates by hand... then have them switch off.
That is essentially what I do and I find it works well. In the
immediately following lecture I show the correspondences between
printouts and calculations (e.g., overheads of printouts with
calculations linked to corresponding output by lines or letters).
> My main goals are to make students realize that stats doesn't have to be
> intimidating, that it is useful, and that they are capable of doing it. Oh,
> and since I'll be seeing most of them again in the fall for the methods
> course, I want to be sure and prepare them with what I feel they should
> know when coming into that course...
Laudable and achievable objectives. Put a lot of weight on
"capable of doing it" and encourage (force?) them to undertake
the considerable work that can be necessary for understanding and
success, as mentioned in my earlier post (e.g., working outside
class, repeated exercises, working in groups, staying on top of
work throughout term, ...).
Best wishes
Jim
============================================================================
James M. Clark (204) 786-9757
Department of Psychology (204) 774-4134 Fax
University of Winnipeg 4L05D
Winnipeg, Manitoba R3B 2E9 [EMAIL PROTECTED]
CANADA http://www.uwinnipeg.ca/~clark
============================================================================
