Correlation Analysis

For more than thirty years scholars have sought to determine to what extent, if any, the Gospel of Thomas used the synoptic gospels as sources. There are roughly 100 sayings attributed to Jesus found in both Thomas and the synoptics. To date no one has found any significant correlation between the order of these paralleled sayings in Thomas and their order in real (e.g. Matthew) or hypotheticial (i.e. Q) canonical texts. [[Robert McL. Wilson wrote, "If numerous sayings in Thomas were identical with those in our Gospels, if they appeared in the same order, or if they appeared in a different order in such a way to suggest that they had been collected under topical heads, there would be grounds for assuming a direct relationship; but in fact it is a notable feature of these sayings that almost all show some variation from the canonical form, while the order is almost entirely different and indications of a topical arrangement are almost entirely lacking," R. McL. Wilson, Studies in the Gospel of Thomas, (London: Mowbray, 1960) 51. Stephen Patterson recently concluded that "a comparison of the order followed by the Gospel of Thomas on the one hand, and the synoptic gospels on the other, has turned up little that would support the theory that the former is dependent upon the latter." 140. S. Patterson, The Gospel of Thomas Within the Develop[ment of Early Christianity, (Claremont Dissertation, 1988).]]

The method used by scholars to investigate the possible correlation in sayings order between Thomas and synoptic gospels may be described as "careful learned consideration," which is not a bad way of going about things. But the question of correlation in sayings order is fundamentally a question for technical statistical analysis, a question which can be answered by computerized analytical programs. This note reports the results of such analysis.

For my collection of Thomas-synoptic parallels I use the recent list compiled by Helmut Koester. As he distinguishes some possible parallels as "doubtful," I compiled two lists, one with and one without "doubtful" parallels. There are 110 sayings in the first set, 100 in the second. I give the results below of the analysis of the set lacking "doubtful" parallels; the results for the other set are not significantly different.

I used the Statistical Analysis System of the SAS Institute Inc. run on a Digital VAX minicomputer to determine the Pearson product-moment correlation. This procedure produces correlation statistics relating the order of sayings in Thomas to the order of sayings in synoptics as well as correlation statistics relating the order of sayings in the synoptics which occur also in Thomas independently of their Thomasine order. In the following Tables the "Degree of Correlation (R)" figure indicates the closeness of linear relationship between the two named variables. The "Reliability of Correlation (p)" number indicates the extent to which the "Degree of Correlation" found was likely to be due to chance factors. Accordingly, the lower the number for "Reliability of Correlation" the lower the likelihood that the relationship was due to chance, and thus the greater the chance that the correlation is significant. Normally if the "Reliability of Correlation (p) exceeds .05 the supposed correlation should be regarded as lacking significance. "N" is the number of pairs of sayings that were analyzed. I will first provide the results of the analysis of correlations between the paralleled sayings in Thomas according to their order in the synoptics.


Table I

Matthew parallels in Thomas to Luke parallels in Thomas
Degree of Correlation (R): .528 N = 62
Reliability of Correlation (p): .0001


Matthew parallels in Thomas to Mark parallels in Thomas
Degree of Correlation (R): .876 N = 29
Reliability of Correlation (p): .0001
Luke parallels in Thomas to Mark parallels in Thomas
Degree of Correlation (R): .709 N = 25
Reliability of Correlation (p): .0001
QMatthew parallels in Thomas to QLuke parallels in Thomas
Degree of Correlation (R): .460 N = 37
Reliability of Correlation (p): .004

The results indicate that the order of sayings in Matthew is more closely correlated to their order in Mark than is the order of sayings in Luke to their order in Mark, and that there is less correlation between the order of sayings in QMatthew and QLuke than there is correlation between either the order of sayings in Matthew or Luke and their order in Mark. In all cases the discovered correlations are not likely to be due to chance. As the sayings correlated here are an essentially random list (i.e. only those happening to occur in Thomas) they are significant primarily to confirm the applicability of the method of correlation adopted here.
Table II

Mark parallels in Thomas' order to Mark's order
Degree of Correlation (R): .213 N = 30
Reliability of Correlation (p): .258


All the Matthew parallels in Thomas' order to Matthew's order
Degree of Correlation (R): .088 N = 82
Reliability of Correlation (p): .431
All the Luke parallels in Thomas' order to Luke's order
Degree of Correlation (R): .141 N = 77
Reliability of Correlation (p): .221
QMatt parallels in Thomas' order to QMatt's order
Degree of Correlation (R): .062 N = 37
Reliability of Correlation (p): .714
QLuke parallels in Thomas' order to QLuke's order
Degree of Correlation (R): .074 N = 37
Reliability of Correlation (p): .663
MkMatthew parallels in Thomas' order to MkMatthew's order
Degree of Correlation (R): .084 N = 25
Reliability of Correlation (p): .689
MkLuke parallels in Thomas' order to MkLuke's order
Degree of Correlation (R): .170 N = 25
Reliability of Correlation (p): .415


No significant or statistical reliable correlations are discovered. Statistical analysis indicates that there is no statistically significant correlation in the order of paralleled sayings in the Gospel of Thomas to the order of sayings in any of the synoptic gospels or to the order of sayings in any source- subset of the synoptic gospels.

Excluding "doubtful" parallels, we have 100 sayings in Thomas which can be compared to the order of 226 other ordered sayings (82 Matt, 77 Luke, 37 Q, 30 Mark). In Thomas we find GT 32,33b // Matt 5:14b-15, GT 43c-45c // QMatt 12:31-35, QLuke 6:43-45, GT 65-66 // Mark 12:31-35 par, GT 92a, 93-94 // Matt. 7:6-7; conceivably these four coincidences of order have text-historical explanations. But four is about the number of such coincidences one would expect through sheer random chance.

Based on observed processes at work in manuscript traditions one might hypothesize that scribes copying Thomas would have been prone to harmonize Thomas sayings order with the order they recalled from one or another of the synoptic gospels, but this evidently did not happen. The lack of correlation in sayings order between Thomas and the synoptics or between Thomas and a source-subset of the synoptics confirms the considered conclusions of previous scholarship and provides evidence both that Thomas did not use the synoptics as a source for sayings.

by Stevan L. Davies


Back To Thomas Homepage