The purpose of this study is to explore the concept of critical worldview integration by looking into a variety of relationships between two areas of worldview belief (religious worldview and epistemological worldview) and several psychological domains (intrinsic/extrinsic motivation, satisfaction with life and vitality among several samples of college students (freshman / sophomores and graduate). This chapter will discuss the results of the data collected in this study.
Samples and Populations
Before any analysis of data can begin, it is necessary to analyze the samples and the population to which results may be generalized. This study essentially uses three populations. The first population is freshman and sophomore students at an evangelical institution in California. This population was sampled using four sections of a required class (426 questionnaires distributed) that included approximately 25 percent of the entire population (1,659 students). Of this sample, approximately 64 percent responded, yielding significant data that allows generalization to that population. The second population consists of students in that same institution’s M.A. Christian Education and M.A. Philosophy and Ethics programs (136 students). Here, 80 questionnaires were distributed with 65 percent responding. This also becomes a sample that may be reasonably generalized to the population of graduate students in these two programs.
However, the third and final population does not exhibit this same integrity. Because the sampling required the use of blind e-mail solicitations to obtain participants for this third sample, little is known about the true population. While the data does indicate that the participants were current freshmen, sophomore and graduate students at California institutions, it does not indicate what caused certain students to respond while others did not. All that is known for certain is that they have e-mail addresses and access to a computer. It is also likely that most watch videos and have an account with Blockbuster™ - although several indicated that they did not want the coupon. Additionally, about a third of those who began the questionnaire failed to complete it, with many quitting shortly after or before the Religious World View Scale. As a result, one might wonder if a certain religious segment tended to drop out at this point. In short, the population lacks definition and results certainly cannot be generalized to a population of college students in California. While it would have been nice to have access to a better defined population, this was a luxury that this study was not afforded. In addition, this study likely underscores the problems involved in doing independent research on multiple campuses and bolsters the notion of cooperative multi-campus, multi-author research in the future. This may serve as a warning to students planning dissertation research.
As a result, the data analysis will focus first on the more generalizable evangelical data and then more tangentially discuss the e-mail invitation or combined data. Due to the obvious exploratory nature of this study, this later data may be helpful within such an exploratory study, but must be treated appropriately. In addition, due to the homogeneous nature of the evangelical data, many issues dealing with differences in Religious World View Scale scores and many of the family upbringing and personal belief questions will not arise until we include and analyze this less anchored e-mail or combined sample. While the analysis of this data may be interesting, it truly provides little other than suggestions for future research.
The discussion of this study’s findings will be organized in the following manner: First, a descriptive analysis of the data collection results is provided. Second, the descriptive results of the demographic information are discussed. Third, this chapter will review the differences resulting from a) sample source (evangelical school versus e-mail invitation); b) gender; c) year-in-school (freshman and sophomores versus graduate students); and d) age (18 to 23 years versus 24 years and up). Fourth, the significant mean differences and correlations relating to the Religious World View Scale (RWVS) scores will be examined. Fifth, any major or unusual correlations discovered between the various scales and their sub-categories will be discussed. The sixth section will specifically examine the important mean differences discovered among the Family Upbringing and Personal Belief Questions. This turned out to be more interesting and informational than anticipated. Finally, the chapter will conclude with a quick summary.
Each variable is generally analyzed based upon scores received from four different instrument groupings. The first is the Religious World View Scale. Note that the participant scores on the Religious World View Scale will be treated as both categorical and interval. When participant scores are divided into five increasing religious worldview categories for purposes of analyzing mean differences, these scores will be used as categorical data. However, for purposes of analyzing correlations between Religious World View Scale (RWVS) scores and other measures, they will be treated as interval data, thereby allowing for such analysis. The second group consists of the Subjective Vitality Scale (SVS) and the Satisfaction with Life Scale (SWLS). Note that each of these scales has a secondary score (SVS2 and SWLS2) that each include one additional question added by the author. SVS2 adds the question “I am energized by a sense of awe and wonder” (Item 72). SWLS2 adds the question “I look forward to learning new things each day” (Item 73). The third group consists of the Epistemic Belief Inventory (EBI) and its five sub-categories. There is no total or cumulative score for the EBI. The final group consists of the Academic Motivation Scale (AMS) and its seven sub-categories. There is also no total or cumulative score for the AMS. Table 7.1 lists each instrument group and its sub-categories in the order in which they are typically reported in the upcoming tables.
Finally, the original null hypotheses and the results are listed below. Note that many of these hypotheses are quite open ended and the different possibilities raised by each will be discussed within the appropriate section.
Effects of Gender: (See Tables 7.9 and 7.10)
H0 1.1: There is no significant difference between gender in any of the five measured dimensions on the Epistemic Beliefs Inventory or the seven types of motivation found on the Academic Motivation Scale. (Rejected).
Effect of family religious upbringing: (See Tables 7.28 through 7.32)
Effect of seeking independent answers: (See Tables 7.33 through 7.36)
Effect of educational level: (See Tables 7.11 and 7.12)
H0 2.1: There is no difference between lower level students and upper level students regarding any of the five measured dimensions on the Epistemic Beliefs Inventory or the seven types of motivation found on the Academic Motivation Scale. (Rejected).
Religious Worldview (See Tables 7.17 through 7.25)
Response Rate and Useable Data
At the evangelical Christian institution, a total of 506 questionnaires were distributed among both undergraduate students and a total of 324 were completed, including 26 pencil and paper questionnaires and 298 that were taken online. At the time that final data was collected, online statistics showed that 47 additional persons started the questionnaire online, but never finished. Any online questionnaires that were not completed (which required all questions to be answered) were discarded. In addition, to the 26 mentioned above, two persons turned in pencil and paper questionnaires that did not answer all questions. That data was also discarded. Finally, these completed responses also included 26 juniors and two seniors. These responses were also not included within the data analysis, leaving a total of 296 evangelical participants. The total response rate including all completed questionnaires was 64%. Table 7.2 shows these statistics in more detail and accounts for differences in the undergraduate and graduate data.
1 Includes unused Junior and Senior responses.
For the participants who were invited by e-mail solicitation, the results were quite different and cannot be compared easily. Here, it was claimed by the e-mail list provider that an anticipated response rate would be around 2 to 3%. Two separate invitations went out to 5,000 random e-mail addresses each. See Table 7.3. The initial invitation resulted in 187 completed questionnaires for a 3.74% response rate. Of these responses, only those for freshmen/sophomores (60) and graduate (101) were used. It is also interesting to note that 276 started the survey, thereby leaving 89 who failed to complete it. Apparently, some became tired, discouraged, or pragmatically decided that it was not worth a free Blockbuster rental coupon. In addition, this campaign resulted in a disproportionate number of graduate students who participated (101).
When the second campaign was launched, it was limited to only those between the ages of 18 and 22 and used an e-mail list that the list provider insisted would be filtered to include only current freshmen and sophomore students within that age group. In addition, the questionnaire was modified to reject responses from those who did not answer that they were current freshmen or sophomores. This campaign yielded another 55 completed freshmen and sophomore responses (only a 1.1% response rate), however there were approximately 50 respondents who opened the survey but were turned away due to not being either current freshmen or sophomores. This would leave the researcher to believe that the list provider was not as accurate in its e-mail targeting as it claimed.
About 60 percent of the combined participants were female. For the evangelical undergraduates the figure increased to 67 percent female, while the evangelical graduate sample dropped to only 23 percent female. See Table 7.4. It was interesting to note that nearly an equal number of males and females commenced taking the online questionnaire; however males had a higher attrition rate. This was true regardless of whether undergraduate or graduate males. The median age was within expectations with a greater range exhibited by the evangelical sample. Within the disclosed field of study, there were no Engineering students in the evangelical sample and few Philosophy students among the e-mail participants. Yet, it must be remembered that the graduate evangelical sample specifically targeted graduate Philosophy and Christian Education students. Otherwise, there was little difference in field of study. See Table 7.5.
Religious Affiliation by RWV Category – Combined Sample
n = 512
Sample Source Differences
It quickly became apparent how very different the evangelical sample was from those who responded through the e-mail invitation. Significant differences between the two samples arose in nearly every category. See Table 7.8. This would suggest that by combining the data from the two sources, the resulting sample would exhibit less difference than if the two samples were compared independently. Here, and throughout these results, t tests for independent samples were used unless extreme differences in variation were noted. In such cases a Welch t test for two samples assuming unequal variance was used and the same is noted within the p value.
St tests: Sample source
As anticipated, the difference in RWVS scores was dramatic (effect size of .40) although it tells us little other than that the e-mail sample really does include a variety of religious worldviews, unlike the homogeneous evangelical sample. Thus, the evangelical institution is effective in capturing students with high degrees of religious orthodoxy. There was a range from 35 to 148 in the e-mail sample scores, while the range was only from 85 to 150 for the evangelical sample. For the e-mail sample the mean RWV score was 98.6, the median was 97.5. The mean score for the evangelical students was 132 with a median of 135. Of more interest is the fact that significant differences arose regarding both Subjective Vitality and Satisfaction with Life with the evangelical sample scoring consistently higher. While it would be nice to conclude that this result was due to either their religious convictions or the quality of the evangelical institution, such causal conclusions cannot be made. This could be explained as simply as wondering if people spending their time on computers answering e-mail solicitations or attempting to garner Blockbuster™ coupons simply tend to be less vital or satisfied with their life.
Regarding epistemic beliefs, the evangelical sample scored significantly lower in Certain Knowledge and Omniscient Authority, while scoring higher in Fixed Ability. Here, the effect sizes were low except for Certain Knowledge which was moderate at .25). This would suggest that the e-mail sample was likely to treat knowledge as more certain in nature. This seems to defy my initial intuition. Whether this result has anything to do with religious worldview will be investigated in the next section.
When it comes to the AMS and motivational issues, there were no significant differences in the Intrinsic sub-categories of Knowledge or Accomplishment. Although differences arose in Intrinsic – Sensation, Extrinsic – Identified, and Extrinsic – Introjected, the effect sizes in each case were quite small. On the other hand, the significant differences in Extrinsic – External Regulation and Amotivation were a bit larger (.07 and .065 respectively). In each case, the evangelical sample scored lower, indicating that they were less susceptible to being amotivated or to being motivated by external coercion.
(Hypotheses H0 1.1 to H0 1.3)
There were no significant differences by gender in the RWVS scores for either the evangelical or the combined (e-mail and evangelical) sample. Nor were there any differences in Satisfaction with Life. However, both samples showed some significant differences (small effect sizes) in Subjective Vitality with females scoring higher. Interestingly, this only became significant for the evangelical sample once the additional item “I am energized by a sense of awe and wonder” was added (Subjective Vitality2). See Tables 7.9 and 7.10.
Regarding EBI scores, there were significant differences by gender in both samples for the sub-categories Certain Knowledge and Fixed Ability, although effect sizes were also small. In both cases, females had higher (allegedly less mature) scores. Finally, there were significant differences by gender in both samples among two of the sub-categories for the AMS; Extrinsic – Introjected (higher for females), and Amotivation (lower for females). An additional significant difference was found in the evangelical sample for Intrinsic – Knowledge (higher for females) and in the combined sample for the sub-categories Intrinsic – Accomplishment and Extrinsic – Identified (both higher for females). However, again all effect sizes were small with Amotivation showing the highest at .06 in the evangelical sample.
t tests: Gender differences – Combined Sample
(Hypotheses H0 2.1 to H0 2.3)
A significant difference appeared between the year-in-school and the RWVS scores (graduates scoring higher) in the evangelical sample only (effect size of .05). This likely suggests only that those with higher RWVS scores elect to continue into graduate programs at the evangelical institution. Neither sample showed any differences in either Subjective Vitality or Satisfaction with Life. See Table 7.11 and Table 7.12.
Regarding EBI scores, there were significant differences between the freshman/ sophomores and graduate students for both samples in the sub-category of Simple Knowledge with graduates scoring higher. The effect size was higher (.05) among the evangelical sample. The combined sample also showed a significant difference in the sub-category of Fixed Ability, with graduates exhibiting more belief that one’s epistemic ability can be improved over time. Here, again, the effect size was quite small. Finally, there were significant differences by gender in both samples among two of the sub-categories for the AMS; Extrinsic – Identified and Extrinsic – Introjected. Graduates scored lower in each, but effect sizes were also low – the evangelical effect size for Extrinsic – Introjected (motivation by internally generated coercion) was the highest (.05). An additional significant difference was found in the evangelical sample for the sub-category of Extrinsic – External Regulation (with graduates scoring lower) and for the combined sample in the sub-category Amotivation with graduates scoring higher. This later result seems a bit surprising. Again, effect sizes are low.
Based upon the forgoing, the null hypothesis (H0 2.1) that there is no difference between lower level students and upper level students regarding any of the five measured dimensions on the Epistemic Beliefs Inventory or the seven types of motivation found on the Academic Motivation Scale is rejected. The null hypothesis (H0 2.2) that there is no difference between lower level students and upper level students regarding their scores on either the Satisfaction with Life Scale or the Subjective Vitality Scale is retained. On the other hand, the null hypothesis (H0 2.3) that there is no difference between lower level students and upper level students regarding their religious worldview scores is retained as to the combined sample, but rejected as to the evangelical sample.
When it came to differences between ages (18 to 23 versus 24 and older), the differences appeared mainly within the evangelical sample. Here, a significant difference appeared between age and the RWVS scores (older students scoring higher) in the evangelical sample only (effect size of .04). Here, again, this likely results from the fact that the graduate students are older and that graduate students who elect to continue their graduate studies at an evangelical institution have higher RWVS scores. Neither sample showed any differences in either Subjective Vitality or Satisfaction with Life based upon age. See Tables 7.13 and 7.14.
The evangelical sample exhibited significant differences based on age between the EBI sub-categories of Simple Knowledge and Quick Learning with the older students scoring higher in both. This matches what was found with graduate students, but both point to allegedly less mature epistemic positions. However effect sizes for each were low. In addition, significant differences based upon age were found among the evangelical students in three sub-categories for the AMS; Intrinsic – Accomplishment, Extrinsic – Introjected, and Extrinsic – External Regulation. In each case, the older students scored lower. While it would make sense that older students would be less coercively extrinsic in their motivation, it seems strange that one’s motivation for accomplishment or creation might go down as one ages. However, once again effect sizes were extremely low. Also, it is interesting to note that the only significant difference for the combined sample was for the sub-category of Extrinsic – Introjected with a minimal effect size.
Differences in Graduate Majors for Evangelical Sample
One question that was raised in the course of analyzing the evangelical data concerned possible differences between the graduates pursing the Masters of Philosophy and Ethics (MA Phil) degree and those pursuing the Masters of Christian Education degree (MACE). Here, although the low number of participants (n = 52) may have prevented significant differences from surfacing, there were no observable significant differences between these groups (Table 7.15). The only two categories that came close to significance were in RWVS scores (where MACE students scored slightly higher) and Intrinsic – Sensation (where MA Phil students scored slightly higher). Although not reaching levels of significance, it was interesting to note that the MACE students had higher EBI scores on Certain Knowledge and Fixed Ability, while the MA Phil students were higher on Simple Knowledge and Quick Learning. The MACE students were also higher on SVS and SWLS scores, although none rose to significance.
Religious World View Scale Categories
(Hypotheses H0 3.1 to H0 3.2)
As expected, the evangelical sample showed a high RWVS score, enough to be concerned that there might be some ceiling effect. In addition, the SD for the evangelical sample (9.71) was much lower than the combined sample (30.237). This is likely a simple correct description of the samples, but it may make it difficult to determine differences within the evangelical sample based upon RWVS groupings. In fact, while the range of scores found in the combined sample allowed for five categories (25 to 50; 51 to 75, 76 to 100, 101 to 125, and 126 to 150), the evangelical sample was only divided into two categories (85 to 125, and 126 to 150).
Dividing the evangelical sample into two categories caused little problem and differences could be calculated using independent t-tests. However, an interesting statistical analysis problem arose once the combined sample was divided up into five RWVS score categories. First, there were extreme differences in the sample size for each category, partly due to the number of high scores from the evangelical sample and partly due to the fact that there were also more Christians than any other religious affiliation in the e-mail sample. See Table 7.16. While designating oneself as a Christian in the demographic portion of the questionnaire did not necessarily result in a high RWVS score (for example the lowest score claimed to be Catholic), Christians did typically score significantly higher. Regardless, there were only 12 scores in the lower category while there were 286 in the upper category.
In addition, possibly due to this sampling issue, the Levene Statistic for several of the ANOVAs showed that the assumption of equal variance had been violated. In these cases, both Tukey’s and Games-Howell (a Tukey’s variation designed for use when equal variances cannot be assumed) post hoc tests were performed. In addition, a Dunnett C post hoc test was often use for further comparison. In nearly every case, the same items remained significant in all three tests, thereby suggesting that there was little computational error due to the unequal variances. However, the sheer magnitude of the difference in sample size still causes concern.
To investigate the effect of this sample size difference, two other methods of assigning RWVS subcategories were used. First, since the e-mail sample (without being combined with the evangelical sample) contained a representative portion of high scores on the RWVS (as well as all of the lower scores), the evangelical sample was removed. This provided for a more equal n distribution among the categories. See Table 7.16. ANOVAs and appropriate post hoc tests were rerun for each of the SVS, SWLS, EBI and AMS sub-categories. Finally, in an attempt to allow for a nearly equal n distribution, a new set of categories was devised for the e-mail sample as shown in Table 7.16. Here, although the range of values within each category now differed, the n was nearly constant at 42 to 45 separate scores. Once again, ANOVAs and appropriate post hoc tests were rerun for each of the SVS, SWLS, EBI and AMS subcategories. As these tests were rerun, a few of the original RWVS category ANOVAs became insignificant. However, the vast majority of the analyses paralleled each other, lending support that the original tests were not as impacted as had been originally feared. The reported analyses for RWVS categories rely upon the e-mail only sample using the original RWVS categories. Any change or notable deviations that are important in the discussion will be mentioned.
A number of significant differences arose between RWVS scores and both Subjective Vitality Scale (SVS) scores and Satisfaction with Life Scale (SWLS) scores. For the evangelical sample, the higher of the two RWVS categories resulted in higher SVS and SWLS scores, but with low effect sizes. See Table 7.17. ANOVAs and resulting post-hoc tests also showed significant differences among the e-mail sample. See Table 7.18. Notice that the effect sizes from the e-mail sample were typically larger than those of the evangelical sample. Additionally, effect sizes for Satisfaction w/ Life 1 and 2 (each .146) were greater that that for Subjective Vitality 1 and 2 (.105 and .107). Recall that “Subjective Vitality 1” is the original SVS, while “Subjective Vitality 2” is the SVS plus a score on the additional question “I am energized by a sense of awe and wonder.” Additionally, that “Satisfaction with Life 1” is the original SWLS, while “Satisfaction with Life 2” is the SWLS plus a score on the additional question “I look forward to learning new things each day.” This suggests that the higher the RWVS score, the higher one’s vitality and satisfaction with life.
(Subjective Vitality, Satisfaction with Life)
Yet, the real area of interest seems to be the patterns exhibited within the e-mail sample data, and not the differences. For all categories, both Subjective Vitality scores reflect that the highest scores were found on the opposite ends of the categories (25 to 50 and 125 to 150) with the lowest score found in the middle category (76 to 100). In addition, the highest score was always found in the 125 to 150 category. See Table 7.19. This curvilinear pattern might suggest that those in the more uncommitted liberal or middle of the road categories did not feel quite as vital. One might argue that to maximize vitality, one should make a decision to be located on one end of the spectrum or the other – with those scoring highest on the RWVS showing the highest vitality levels. This same pattern is true for Satisfaction with Life scores with one major exception. Despite an increase in SWLS scores as RWVS scores lowered from category 76 to 100 to category 51 to 75, the SWLS scores then dramatically drop to their lowest value in the RWVS category 25 to 50. Here, unlike with vitality, the lowest RWVS scores also have the lowest satisfaction with life scores. Figure 4 is a line chart of these mean scores and depicts this pattern graphically. These patterns clearly warrant further investigation.
Within the evangelical sample, there was a significa Knowledge scores increased, while both Certain Knowledge and Omniscient Authority scores decreased. Effect sizes in each case were low, however Certain Knowledge approached .05.
Knowledge scores increased, while both Certain Knowledge and Omniscient Authority scores decreased. Effect sizes in each case were low, however Certain Knowledge approached .05.nt difference between RWVS scores in three EBI sub-categories: Simple Knowledge, Certain Knowledge and Omniscient Authority. See Table 7.20. Here, as RWVS scores increased, Simple
n = 216 * = lowest score
Figure 4. Plot of mean Subjective Vitality and Satisfaction w/ Life scores for each Religious World View Scale category. (E-mail Sample).
(Epistemic Beliefs Inventory)
One way Anovas for the five categories in the e-mail sample showed significant differences for each EBI sub-category. Effect sizes ranged between .049 to .119 for all categories except “Quick Learning” – which showed a much higher effect size of .212. See Table 7.21. Post hoc tests confirmed significant mean differences between each RWVS category except in “Fixed Ability” where the differences between categories were not significant after performing a Games-Howell test. General patterns were once again interesting. See Table 7.22. The pattern for both Simple Knowledge and Quick Learning is similar to that of SVS, but less pronounced. Both the lowest and highest categories find knowledge to be more simple and a quick all or none commodity, while the middle category finds knowledge to be more complex and capable of being learned over time (the claimed more mature position). See Figure 5 for a line chart showing these patterns.
The pattern for Omniscient Authority exhibits EBI scores that decrease as RWVS scores increase. Thus, interestingly, the higher the RWVS score, the less knowledge is dependent on authority (the claimed more mature position). This is the opposite of what was expected. One interesting possibility is that the EBI questions speak about men as authorities and never mention God as a possible authority for either creating or dispensing knowledge. Thus, those with higher RWVS scores may only consider the authority of man and agree that we certainly cannot rely on man as that authority – a position of humility.
n = 216 * = lowest score
Figure 5 Plot of the mean scores for each factor of the Epistemic Belief Inventory for each Religious World View Scale category. (E-mail Sample).
Significant differences were also found between RWVS scores in the evangelical sample in three AMS sub-categories. With Intrinsic – Knowledge (effect size .06), motivation for the intrinsic sake of gaining knowledge increased with increasing RWVS scores. On the other hand, Extrinsic – External Regulation and Amotivation (each with effect sizes of approximately .03) dropped as RWVS scores increased. See Table 7.23.
Anovas conducted with the e-mail sample also showed significant differences between RWVS scores and the sub-categories of Intrinsic – Knowledge, Extrinsic – External Regulation, and Amotivation. While most of the effect sizes ranged around .07 and .08, the effect size for Amotivation was substantially larger at nearly .20. Unlike the evangelical sample, significant differences were also found in the sub-categories of Intrinsic – Sensation and Extrinsic – Identified. See Table 7.24.
In the case of Extrinsic – External Regulated, the pattern showed that generally the higher the RWVS score, the lower the AMS sub-category score. See Table 7.25. See also Figure 6 for a line chart showing these patterns for the Intrinsic sub-categories and Figure 7 for a line chart showing the patterns for the Extrinsic sub-categories. Thus, those with lower RWVS scores tend to find themselves more motivated by external coercion, threat or pressure, while those with higher RWVS scores do not. In the case of Amotivation, the pattern was generally similar to that found for vitality. Both the highest and lowest RWVS scores were least likely to be amotivated. Those in the middle category (76 to 100) were significantly more likely to operate under amotivation – without any real purpose for their actions or motivations.
* lowest score
Figure 6 Plot of the mean scores for the Intrinsic Motivation factors of the Academic Motivation Scale for each Religious World View Scale category. (E-mail Sample).
Figure 7 Plot of the mean scores for the Extrinsic Motivation and Amotivation factors of the Academic Motivation Scale for each Religious World View Scale category. (E-mail Sample).
This same pattern was found for the sub-category Intrinsic – Knowledge, although not as pronounced. Here, those with RWVS scores in the low category were more likely to be motivated by an intrinsic desire to simply acquire knowledge, while those in the middle and high categories exhibited lower levels. This suggests that while those with low RWVS scores tend to be more motivated by coercion, they also tend to be more motivated by a simple quest for knowledge. Due to difficulties in generalizing beyond the evangelical population, this certain warrants further investigation. It was interesting to note that these patterns became even more pronounced when running the same ANOVA and post hoc tests on the combined sample rather than just the e-mail sample.
Based upon the forgoing, the null hypothesis (H0 3.1) that there is no significant difference between those falling in different religious worldview categories regarding any of the five measured dimensions on the Epistemic Beliefs Inventory or the seven types of motivation found on the Academic Motivation Scale is rejected. Likewise, the null hypothesis (H0 3.2) that there is no significant difference between those falling in different religious worldview categories regarding their scores on either the Satisfaction with Life Scale or the Subjective Vitality Scale is also rejected.
Copyright 2005 Mark E. Henze - BBuilders.org