Table 1

Guide to Understanding the Article

Section

 Tips, Comments, and Problem Areas

Abstract

Buffers: protects; shields

Devout intrinsic religion: feeling deeply devoted to a set of religious beliefs

Introduction

Top of page 327

As discussed in Chapter 2, Chapter 6, Chapter 8, Chapter 9, and elsewhere in Research design explained, correlational research does not allow one to make cause-effect statements.

The Present Research

Note that the authors’ hypothesis is an example of using two principles we discuss in Chapter 2 of Research design explained: Principle 4 (page 50) and Principle 10 (page 52).

Method

Today, we would call the students volunteering for the study “participants” rather than “subjects.”

Randomized block procedure: a technique to randomly assign participants to condition while making sure that an equal number of participants are assigned to each group.

Note that the present study is an experiment because the researcher randomly assigned participants to condition.

Procedure

Note that the researchers invented neither their measures nor their manipulation. Instead, by searching the literature, the authors were able to find measures they could use and a manipulation that was both effective and ethical.

Debriefing

Note two ethical benefits of debriefing. First, debriefing allowed the researchers to detect and remedy any harm the manipulation caused students. Second, debriefing allowed the researchers to exclude data that were questionable due to participants’ suspicions.

Results and Discussion

t < 1:  ts below 1 will not be statistically significant (reliable).

Change in intrinsic scores.

 You can do all the analyses repeated in this section by plugging in the relevant values from Table 1 into the spreadsheet at

http://glass.ed.asu.edu/stats/analysis/t2test.html

Note that because of rounding error, your results—although close to the authors’—may not be an exact match.

Footnote 2

Although some of the information in this footnote is highly technical, you can easily understand parts of it. If you read Chapters 8 and 9 of Research design explained, you would probably agree with the authors that, because of random assignment, regression should not have been a serious threat to the study’s validity. That is, that random assignment made the groups similar, there was no reason to expect that the high salience group would be more vulnerable to regression than the low salience group would be. You may have noted another implication of the researchers using random assignment: The researchers could have just administered the manipulation followed by the measure. That is, they did not have to administer the pretest and calculate a change score. The reason the researchers used the pretest was to have more power. This approach paid off because if they had used the posttest scores, they would not have obtained significant results (as you can see for yourself by plugging in the posttest score means, standard deviations, and sample sizes at http://glass.ed.asu.edu/stats/analysis/t2test.html). However, the researchers did not have to calculate change scores. There are two other ways the researchers could have used the pretest scores and boosted power without calculating change scores. First, as the footnote mentions, the researchers could have controlled for individual differences in pretest scores by using a technique that combines correlation/regression with analysis of variance called analysis of covariance (ANCOVA). Second, the authors could have used a 2 (low salience/high salience) X 2 (pretest versus posttest) mixed analysis of variance (to learn about mixed analysis of variance, see Chapter 12 of Research design explained).

Change in extrinsic and quest scores. The authors could have done a more direct test of whether the manipulation had a difference effect on intrinsic religion than on other scales by doing a 2 (low salience/high salience) X 3 (intrinsic religion/extrinsic religion/quest) mixed ANOVA. If the manipulation had a differential effect on intrinsic motivation, there would be a salience by measure interaction[1]. To learn more about mixed analysis of variance, see Chapter 12 of Research design explained.)

Effect on Intrinsic Religion Depending on Resolution of the Loneliness Experience Described.

The two explanations were (a) Freud’s explanation that feeling lonely made one need God and (b) writing about loneliness just made one reflect more on how wonderful God is.

Serendipitous: unintentional consequence

Cronbach’s alpha: Normally, researchers do not use Cronbach alpha as an index of inter-rater reliability. The preferred index is Cohen’s kappa. For more on inter-rater reliability, go to the following site: http://www.temple.edu/mmc/reliability/#What%20is%20this%20site%20about

Note that many statisticians would call p < .06, not significant.

Is Intrinsic Religion Really Intrinsic?

Bestowed: given, awarded

Functionally autonomous: independent of; not the result of; not affected by; free from the influence of

Extrinsically: not of its own sake; affected by outside influences

Viable: workable; defensible; can survive; will not be immediately destroyed

Is the Intrinsic Scale a Measure of Intrinsic Religion?

Idiographic: looking at for principles that apply to individual cases and by viewing each individual as unique, not looking for general principles that apply to most people.

Nomothetic: looking for general principles and laws. Most psychological research searches for general principles of behavior.

References

This reference section is not in APA style.