Page 376 - Statistics for Dummies
P. 376
360
Statistics For Dummies, 2nd Edition
descriptive statistics (continued)
experiments, 272–273
median, 73–74
overview, 14–15, 67–68
surveys, 253
ethical review boards (ERBs), 253, 272
percentiles, 83–89
range, 80
evaluating
bar graphs, 102
standard deviation, 77–80
confidence intervals, 213
in summaries, 14–15
designing studies to answer a research
histograms, 110–112
pie charts, 97
question
time charts, 127
with experiments, 12
overview, 11
exaggerations. See misleading statistics
exam strategies
with surveys, 11–12
breaking down problems, 343–344
detecting errors, 34
discrete data, 46
double-checking your work, 344
formulas, being comfortable using, 336–337
discrete random variables, 132–133, 143.
See also binomials
“if-then-how” chart, making an, 337–339
“know what you don’t know, and then do
distribution. See also specifi c distributions
conditional, 305–308 ethical issues
something about it,” 332
described, 164, 299 labeling everything in problem, 340–342
joint, 302–305 overview, 333
marginal, 299–302 picture form, expressing your exam
overview, 16–17, 54–55 question in, 342–343
two-way tables, 299–308 problems, understanding what the
double-blind experiments, 57–58, 274 question is asking in test, 339–340
double-checking your work as exam real exam conditions, practicing under,
strategy, 344 334–336
sense, analyzing your answers to confi rm
• E • they make, 344–345
“yeah-yeah” traps, avoiding, 333–336
empirical rule (68-95-99.7), 53, 81–83, 112 on your own, trying problems, 334
erroneous data, removing, 41 examples
errors media and statistics, 24–30
in arithmetic, 34 misleading graphs, 322
described, 34, 166, 225 normal distribution, 144
detecting, 34 regression line, 291–292
erroneous data, removing, 41 excess data in time charts, 126–127
in interpretation of survey results, 259–260 experiments
in interpreting boxplots, 122 analyzing data from, 274–277
missing data as factor in, 41 bias in, minimizing, 13–14
of omission, 34 blind, 57
reasonableness of projections, comparisons, defining the experiment to
examining, 34 make, 265–267
type-1 errors (false alarms), 225–226 confounding variables, 270–271
type-2 errors (missed detection), 226 control group, 57, 266
estimates, 32, 193–194. See also confi dence data collection for, 273–274
intervals
3/25/11 8:12 PM
32_9780470911082-bindex.indd 360
32_9780470911082-bindex.indd 360 3/25/11 8:12 PM
