Limitations Due to the voluntary nature of the elicitation process and to the lack of
definition of the group from which volunteers were selected, the results from
this survey can not be used to conduct power analysis in preparation for
larger surveys. Also, inferences concerning larger groups and sub-populations
can not be supported. For example, within the population of respondees, we
have designations for denominational affiliation; for example, Roman Catholic
or Presbyterian. It might be tempting to think that we can say something
significant about people's attitudes toward evolution and Christianity from
these groups based on the results of this survey. This is not warranted,
however. To illustrate this point, say that we want to run some t tests to
look for significant differences in the means between sub-populations
designated by religious denomination, age, or education level. We can run
t-tests on the entire data set or on a fraction of the data set restricted to
one or two modes of collection. Data for the survey was collected in four
different manners: direct elicitation from some SPU students; email to
associates; email to science-religion interest groups; and elicitation from
the counterbalance web-site. Looking at t-tests run on the entire data set as
opposed to a partial data set which removes most responses from the
counterbalance web-site shows so much variability that it appears that the
methods of collecting data are biasing our observations in ways that we are
not in a position to understand relative to the categories of interest.
|
alpha values from t tests |
| Q |
Part. |
Full |
Method |
| Q5 |
0.1166 |
0.7703 |
0.039** |
| Q8 |
0.0099* |
0.0052* |
0.230 |
| Q9 |
0.6415 |
0.1845 |
0.028** |
| Q11a |
0.3222 |
0.1384 |
0.140 |
| Q11b |
0.3603 |
0.5547 |
0.330 |
| Q11e |
0.1438 |
0.1333 |
0.360 |
| Q11i |
0.6462 |
0.5377 |
0.300 |
| Q11l |
0.4089 |
0.9423 |
0.990** |
* significant for test of inequality for alpha = .05
** significant for test of inequality for alpha = .1.
Column 1 gives the question from the survey which is being considered.
Columns 2 and 3 test for significant differences in means between responses,
to various questions, from Catholics or Presbyterians. Column 2 is based on
data primarily from email and SPU sources, while column 3 also includes data
from the Counterbalance web-site. Column 4 gives significance results from
Catholics who participated via Counterbalance versus Catholics who
participated through one of the other elicitation channels.
For what it is worth, within the population of respondees, there is a
significant difference in the means of people's attitudes toward the
compatibility of evolutionary theory and Christian theology depending on
whether they identify as Roman Catholic or as Presbyterian.
Email
link | Printer-friendly | Feedback
| Contributed by: David Caccia
|
