When people read, hear, or prepare research summaries,
they sometimes have misconceptions about what is or isn't "sound
practice" regarding the collection, analysis, and interpretation
of data. Here are some of these common (and dangerous) misconceptions
associated with the content of Chapter 5.
Lengthy, detailed descriptions of samples are not
only boring but actually bad because they detract from the most important
part of the research report, the results.
If data are collected from each and every person
the researcher can possible use in his/her investigation, no inferential
statistics should be used.
A large sample that's not random is superior to a
small sample that is random.
In order to extract a stratified random sample from
a population, the population must be subdivided so as to create strata
that are the same size.
If a researcher ends up with a response rate for
a particular mailed survey or questionnaire that's higher than the "typical"
response rate reported in authoritative texts written about research
methodology, then he/she is entitled to feel pleased about his/her response
Statistical inference can only be used in conjunction
with probability samples.
You can determine whether or not a sample was selected
randomly from a tangible population by looking to see if the sample
is "odd" in some way.
All samples are random samples.
In studies having a less than desirable response
rate to a mailed questionnaire or survey, the best way to check to see
if there's a "response bias" is to compare "early returns"
vs. "late returns."
A study cannot yield interesting, valuable, and generalizable
findings unless its findings are derived from random samples.