Chapter 4

RESULTS

4.0 Introduction

This chapter presents a description of the data collected and the analysis performed to explore the research questions. Descriptive statistics and the appropriate statistical analysis used to compare amongst the actors, meditators and their control groups on their performance in the CANTAB tests, will be presented. Finally the relationship between the total amount of practice in the actors and meditators and their performance in the tests will be explored.

4.1 Descriptive Statistics

Two independent samples with their respective paired control groups participated in the experiment. The control groups were matched on the variables of age, gender and level of education. The total sample was of 48 participants. Each group had 12 participants, constituting 25% of the sample. The gender and age characteristics of each group are summarised in Table 1.

Table 1 Gender and age characteristics of the sample

12.5% of the total sample had a secondary level of education, 8.33% a post-secondary level of education, 37.5% were students at a tertiary level, and 45.83% were graduates in professional employment.

Figure 4 Graph showing the level of education of participants

The mean total hours of training in the actors and meditators population (n = 24) was 3605.531 hours (SD = 4035.781). The mean total training hours for the sample of actors (n = 12) was 1135.333 hours (SD = 1126.861). The range of total training hours in actors varied from a minimum of 312 hours to a maximum of 4160 hours.

A frequency distribution of the total training hours shows that 66.7% of the actors have trained between 312 and 1000 hours. 16.6% have trained between 1000 and 2000 hours, whilst 16.6% have trained between 2000 and 4160 hours.

The mean total training hours for the meditators group (n = 12) was 6075.729 hours (SD = 4412.793). The range of meditation hours varied between a minimum of 730 hours and a maximum of 16060 hours. A frequency distribution of the total meditation hours shows that 66.7% of the meditators have practised meditation between 730 and 6000 hours. 16.6% have practised between 6000 and 10500 hours, whilst 16.6% have meditated between 10500 and 16500 hours.

Figure 6 Histogram showing the frequency distribution of total meditation hours in meditators

4.2 Comparisons between Samples

This section deals with the differences in performance on CANTAB tests among the four groups. For each set of scores obtained on the tests the descriptive statistics will be outlined, followed by a general analysis of variance (one-way ANOVA), and a series of planned comparisons.

The main data items for the IED were the stage reached, the total errors, errors at ED-shift, and errors up to ED-shift. The score, standard score and percentile of each of these data items were automatically recorded by the CANTAB.

The mean stage reached by the total population (n = 48) was 7.792 (SD = 1.935), with a minimum score of 1 and a maximum of 9. Table 2 illustrates a comparison of the mean stage reached by the four groups.

Table 2 Comparison among groups on stage reached in IED

Figure 7 Graph of the stage reached in IED

Figure 8 Graph of the stage reached in IED (Std-Score)

A one-way ANOVA showed no significant differences (F (3, 44) = .204, p > .05) among the four groups on the stage reached in the IED.

Any significant differences between the actors and meditators groups were further explored using a t-test for independent samples. No significant difference was found between the actors and meditators on the stage reached in the IED (t (22) = .104, p > .05). Further analysis was performed between the actors and their control group using a paired t-test. No significant difference was found between the pair (t (11) = .115, p > .05). A paired t-test was also performed between the meditators and their controls. Once again the difference was not significant (t (11) = .609, p > .05).

The whole population (n = 48) had a mean of 24.980 (SD = 11.635) on total errors made in the IED, with a minimum of 7 errors and a maximum of 54. Table 3 illustrates a comparison of the mean total errors among the four groups.

Table 3 Comparison among groups on total errors in IED

Figure 9 Graph of the total errors in IED

Figure 10 Graph of the total errors in IED (Std-Score)

A one-way ANOVA showed no significant differences (F (3, 44) = .082, p > .05) among the four groups on the total errors in the IED.

Differences between the actors and meditators groups were further explored using an independent t-test. No significant difference was found between the actors and meditators on the total errors in the IED (t (22) = -.414, p > .05). A paired t-test, performed between the actors and their control group also showed no significant difference (t (11) = -.275, p > .05). Similarly no significant differences were evident between the total errors of the meditators and their controls when tested with a t-test for paired variables (t (11) = .155, p > .05).

The mean amount of errors at ED-shift of the whole population (n = 48) was 12.625 (SD = 10.977), ranging from a minimum of 0 to a maximum of 31. Table 4 illustrates a comparison of the mean errors at ED-shift among the four groups.

Table 4 Comparison among groups on errors at ED-shift in IED

Figure 11 Graph of the errors at ED-shift in IED

Figure 12 Graph of the errors at ED-shift in IED (Std-Score)

A one-way ANOVA showed no significant differences (F (3, 44) = .823, p > .05) among the four groups on errors at ED-shift in the IED.

An independent t-test between the actors and meditators also showed no significant difference on errors at ED-shift in the IED (t (22) = .062, p > .05). Further analysis was performed between the actors and their control group using a paired t-test. No significant difference was found between the pair (t (11) = -.886, p > .05). A similar paired t-test between the meditators and their controls gave no significant difference either (t (11) = .258, p > .05).

The mean of the whole population (n = 48) on errors up to ED-shift in the IED was 7.479 (SD = 5.520), ranging from a minimum of 0 to a maximum of 28. Table 5 illustrates a comparison of the mean errors up to ED-shift among the four groups.

Table 5 Comparison among groups on errors up to ED-shift in IED

 Figure 13 Graph of the errors up to ED-shift in IED

Figure 14 Graph of the errors up to ED-shift in IED (Std-Score)

A one-way ANOVA showed no significant difference (F (3, 44) = 1.310, p > .05) among the four groups on the errors up to ED-shift in the IED.

The difference between the actors and meditators groups was further explored using a t-test for independent samples. An almost significant difference was found between the actors and meditators on the errors up to ED-shift in the IED (t (22) = -1.967, p > .05). Further analysis between the actors and their control group using a paired t-test also showed an almost significant difference between the pair (t (11) = -2.119, p > .05). A paired t-test was also performed between the meditators and their controls. Any difference was not significant (t (11) = .854, p > .05).

The most significant data items for the RVP were the probability of hit, and the probability of false alarm. The score, standard score and percentile of these data items were automatically recorded and computed by the CANTAB.

The mean probability of hit of the total population (n = 48) was .687 (SD = .187), with a minimum score of .3 and a maximum of 1. Table 6 illustrates a comparison of the probability of hit of the four groups.

Table 6 Comparison among groups on probability of hit in RVP

Figure 15 Graph of the probability of hit in RVP

Figure 16 Graph of the probability of hit in RVP (Std-Score)

A one-way ANOVA showed no significant differences (F (3, 44) = .366, p > .05) among the four groups on probability of hit in the RVP.

Any significant differences between the actors and meditators groups were further explored using a t-test for independent samples. No significant difference was found between the actors and meditators on probability of hit in the RVP (t (22) = -.273, p > .05). Further analysis was performed between the actors and their control group using a paired t-test. No significant difference was found between the pair (t (11) = -.851, p > .05). A paired t-test was also performed between the meditators and their controls. Once again the difference was not significant (t (11) = .514, p > .05).

4.2.2.2 Probability of False Alarm

Table 7 Comparison among groups on probability of false alarm in RVP

Figure 17 Graph of probability of false alarm in RVP

Figure 18 Graph of probability of false alarm in RVP (Std-Score)

A one-way ANOVA showed no significant difference (F (3, 44) = 1.488, p > .05) among the four groups on probability of false alarm in RVP.

The difference between the actors and meditators was further explored using a t-test for independent samples. A significant difference was found between the actors and meditators on the probability of hit in RVP (t (22) = -2.419, p < .05). Nevertheless a further analysis between the actors and their control group using a paired t-test showed no significant difference between the pair (t (11) = -1.393, p > .05). Neither did a paired t-test performed between the meditators and their controls. Any difference was not significant (t (11) = .321, p > .05).

The most significant data items for the PAL were the total trials and mean errors made. The scores, standard scores and percentiles of these data items were automatically recorded and computed by the CANTAB.

The total sample (n = 48) had a mean of 10.938 (SD = 2.513), which ranged from a minimum score of 8 and a maximum of 20, on total trials in the PAL. Table 8 illustrates a comparison of total trials amongst the four groups.

Table 8 Comparison among groups on total trials in PAL

Figure 19 Graph of total trials in PAL

A one-way ANOVA showed no significant differences (F (3, 44) = 2.300, p > .05) among the four groups on total trials in the PAL.

Any significant differences between the actors and meditators groups were further explored using a t-test for independent samples. No significant difference was found between the actors and meditators on total trials in PAL (t (22) = .091, p > .05). Further analysis was performed between the actors and their control group using a paired t-test. No significant difference was found between the pair (t (11) = .508, p > .05). However analysis using a t-test for paired variables, revealed a significant difference between the meditators and their controls on total trials in PAL (t (11) = -2.232, p < .05).

The mean of the whole population (n = 48) on average errors in the PAL was 1.038 (SD = 1.048), ranging from a minimum of .000 to a maximum of 5.400. Table 9 illustrates a comparison of the mean errors at ED-shift among the four groups.

Table 9 Comparison among groups on mean errors in PAL

Figure 21 Graph of the mean errors in PAL

Figure 22 Graph of mean errors in PAL (Std-Score)

A one-way ANOVA showed that there were significant differences (F (3, 44) = 3.304, p < .05) among the four groups on mean errors in PAL. Post-hoc testing using the LSD and Tukey HSD tests (significance level at .05) showed that there were differences between the meditators and their controls.

An independent t-test showed no significant difference between the actors and meditators on mean errors in PAL (t (22) = .139, p > .05) and a paired t-test between the actors and their control group also yielded no significant difference (t (11) = .550, p > .05). However a significant difference was evident between the meditators and their controls when tested with a paired t-test (t (11) = -2.704, p < .05).

The most significant data items for the SWM test were the between errors, and the strategy score. The score, standard score and percentile of these data items were automatically recorded and computed by the CANTAB.

The mean between errors in SWM of the total population (n = 48) was 17.583 (SD = 18.211), with a minimum score of 0 and a maximum of 68. Table 10 illustrates a comparison among the four groups on between errors in SWM.

Table 10 Comparison among groups on between errors in SWM

Figure 23 Graph of between errors in SWM

Figure 24 Graph of between errors in SWM (Std-Score)

A one-way ANOVA showed no significant differences (F (3, 44) = .411, p > .05) among the four groups on between errors in SWM.

Any significant differences between the actors and meditators groups were further explored using a t-test for independent samples. None of these were significant (t (22) = -1.000, p > .05). Further analysis was performed between the actors and their control group using a paired t-test. No significant difference was found between the pair (t (11) = -.163, p > .05). A paired t-test was also performed between the meditators and their controls. Once again the difference was not significant (t (11) = .629, p > .05).

The mean of the whole population (n = 48) on strategy score in SWM was 32.188 (SD = 4.423), ranging from a minimum of 21 to a maximum of 39. Table 11 illustrates a comparison of the strategy score among the four groups.

Table 11 Comparison among groups strategy score in SWM

Figure 25 Graph of strategy score in SWM

Figure 26 Graph of strategy score in SWM (Std-Score)

A one-way ANOVA showed no significant difference (F (3, 44) = 1.479, p > .05) among the four groups on strategy score in SWM.

A t-test for independent samples also found no significant difference between the actors and meditators (t (22) = -1.261, p > .05). Further analysis between the actors and their control group using a paired t-test also showed no significant difference between the pair (t (11) = .631, p > .05). However a paired t-test performed between the meditators and their controls, almost reached significance (t (11) = 1.960, p > .05).

4.3 Total training hours

An analysis of the relationship between the total amount of practice for actors and meditators, and the performance on the CANTAB tests was performed. No relationship was found between the total amount of training in actors and their performance on the IED, PAL and SWM. However there was a strong positive correlation between the total hours training in actors and their probability of false alarms in the RVP (r = .833, p < .01). No significant correlations were found between the total hours practising meditation and the meditators' performance on the IED, RVP, PAL and SWM tests. Tables 12 and 13 will illustrate the results obtained using Pearson's two-tailed correlations for actors and meditators.

Table 12 Pearson's correlations between total hours training in actors and performance on CANTAB tests

** p < .01

Table 13 Pearson's correlations between total hours practising in meditators and performance on CANTAB tests