|
Quantitative and Qualitative Research
The "Data Collection Procedures," which are the "step-by-step
action plan" for how you will collect your information. This would
be something like the following: "On Friday, May 27, 1998, the researcher
and two assistants went to School X to administer the XYZ Test of Motivational
Level. They administered the test to two classes of 30 fourth graders
apiece. The test took 2 1/2 hours to complete." Notice how you are sort
of documenting, "journalism class" style, the who-what-when-where of the
process of data collection ... so that an outside reader could picture
and hopefully reproduce all the steps you took to gather your data.
The "Data Analysis Procedures: "ah, the 'fun stuff!" This describes
how you compiled and analyzed your data, or information, to answer your
research question(s)!
We usually think of statistical procedures, but it's only one way, as you will see below.
For quantitative data (information in numbers), you could have:
- Summary Descriptive Statistics: You sort of "tally
up" and present as a "numeric profile:" e.g., means, totals, percentages,
and the like. You don't 'test' these: they are just meant to give a
broad general summary numeric picture of your data. For instance: how
many (total and percent) of subjects in your sample were men and how
many were women.
- Inferential (also called Analytic Statistics): would be the "testables," some of which we'll be learning in Intro
to Statistics! For instance, say that you find an average difference
in science aptitude test scores between the sixth-grade boys and girls
IN YOUR SAMPLE. But say, also, that you want to use this SAMPLE result
to see if the same thing would be true FOR THE ENTIRE POPULATION of
sixth grade boys and girls in the state (from which you drew this sample
and to which you wish to 'project' or 'generalize' the findings/results).
Could your sample have been 'just a flukey one?' That is: you do get
an average science difference in your sample, but it would NOT hold
up in the POPULATION AT LARGE? or is your sample quite representative
of this population, so that you CAN confidently predict that the SAMPLE
difference in average science aptitude WOULD hold up in the POPULATION
AT LARGE? You'd be using an "inferential" or "analytic" statistic (in
this case, something called an "independent-samples t-test" -- stay
tuned in Intro to Statistics! It'll be coming) to make this decision
about "what to assume about the population based on sample results"
But there's another form that data, or information, can take! What if
these data are NOT in numbers, but in the form OF WORDS -- that is,
qualitative?!
If you have qualitative data, you would also describe, under Data Analysis
Procedures, how you will summarize and compile these data to address your
research question!
There is a third possibility: why not have the best of BOTH
worlds? and collect both numeric (quantitative) and verbal (qualitative)
data to address your problem statement?!
This is an exciting new direction of research called "multimethod
research!"
Having the data both ways in essence is like a "2nd medical opinion,"
or "2 pieces of evidence (favoring innocence of client) in a jury trial!"
With more than one piece of evidence, you can have greater confidence
in the "answer" to your research question(s) if both the numbers
and words appear to "point in the same direction!" This is called
"convergence" or "triangulation" of the alternative data
sources!
Example: You wish to identify the general direction and level
of attitudes towards a new teacher incentive program in your school district.
You decide to collect the data 2 ways:
- having the teacher-subjects rate a group of items reflecting their
attitude on a 1-5 scale; AND
- interviewing a small sample of teacher-subjects, letting them express,
in their own words and in response to open-ended interview questions,
their attitudes.
You compare the above 2 sets of results and find:
- the teachers who completed the rating scale items on the attitude
survey are, on average, giving "moderately favorable" and "very favorable"
attitude ratings to the items; AND
- in the interview sessions, the interview subjects are choosing
to voice primarily positive attitudes in the opinions they share with
the interviewer.
Thus, you can have greater confidence that "attitudes towards the teacher
incentive plan appear to be generally favorable" with the above two sources
converging, than you would if you had only collected data using a single
source! That's the beauty of multimethod research designs! (commercial
announcement: I teach a graduate course on Multimethod Research Procedures
too! It would be an ideal course by modem candidate for us for the future
if you'd be interested!!!)
Once you have finished, you should:
Go on to Assignment 1: Quiz: Quantitative
and Qualitative Research
or
Go back to Collecting and Analyzing Data
E-mail M. Dereshiwsky at statcatmd@aol.com
Call M. Dereshiwsky at (520) 523-1892
Copyright 1998 Northern Arizona
University
ALL RIGHTS RESERVED
|