### Table 1: Rubric used for coding use of sufficient evidence. Causal Element Sufficient data to support claim

2001

"... In PAGE 7: ... To create a persuasive explanation, of course, requires not just relevant data, but sufficient relevant data to warrant a claim. To analyze sufficiency of data, we created a general rubric for each of the four causal elements of a natural selection explanation ( Table1 ). Note that this judgment of sufficient data underlies the judgment of inferential validity in the overall quality score.... ..."

Cited by 1

### Table 1: Rubric used for coding use of sufficient evidence. Causal Element Sufficient data to support claim

2001

"... In PAGE 7: ... To create a persuasive explanation, of course, requires not just relevant data, but sufficient relevant data to warrant a claim. To analyze sufficiency of data, we created a general rubric for each of the four causal elements of a natural selection explanation ( Table1 ). Note that this judgment of sufficient data underlies the judgment of inferential validity in the overall quality score.... ..."

Cited by 1

### Table 3. Centralized data are compared to collective data with time windows 150, 180, 210, 240, 270 and 300 seconds. Table lists resulting P values of Kolmogorov-Smirnov Goodness-of-fit tests (kstest) for interarrival time and message size (in word count). P values: (i) gt; 0.05 mean difference between two data sets is statistically insignificant, (ii) 0.01 to 0.05 mean difference is significant, (iii) 0.001 to 0.01 mean difference is very significant, and (iv) lt; 0.001 mean difference is extremely significant. P values should be gt; 0.05 to be able to conclude that there is no sufficient evidence to reject the hypothesis that two data sets are coming from the same distribution. For interarrival time, P values gt; 0.05 for the servers S1, S2, S4 and S8 which have highest message counts as given in Table 1. There is no sufficient evidence to reject that interarrival time distributions of these pairs of data sets are the same. For message size, all P values are gt; 0.05, meaning that, difference between two data sets is statistically insignificant and there is no sufficient evidence to reject that message size distributions of these pairs of data sets are the same.

2006

"... In PAGE 4: ... These statistics support findings in [1] that interarrival and message size fit to exponential distributions. Table3 provides results of statistical com- parison between collective and centralized data based on Kolmogorov-Smirnov goodness-of-fit test (kstest). Centralized data are compared to collective data in terms of interarrival time and message size distributions.... ..."

Cited by 4

### Table I summarizies the data for the 64 peptides studied. Here we present evidence that the simulations were sufficiently equilibrated to draw broad conclusions over groups of peptides. Then we show that the observed helicities correlate with I-sites confidence values as well as with AGADIR helix propensity values. We also show that capping the ends of 8-residue peptides increases helicity.

### Table 2. Assessing the strength of a body of evidence on effectiveness of population-based interventions in the Guide to Community Preventive Services

"... In PAGE 5: ... Depending on appropriateness and feasibility of a quantitative summary and the availability of statistical measures of variability or data from which to calculate them, formal procedures for statistical pooling might also be used to describe a summary measure of effect. In addition, the body of evidence of effectiveness is characterized as strong, sufficient, or insufficient based on the number of available studies, the strength of their design and execution, and the size and consistency of reported effects ( Table2 ). Sufficient and strong evi- dence can be achieved in several ways that incorporate scientific rigor and the feasibility and appropriateness of evaluation for the wide range of interventions used in population-based approaches to improve health.... In PAGE 5: ... Several principles guided the designation of bodies of evidence of effectiveness as strong, sufficient, or insufficient evidence. Strong or sufficient evidence can be based either on a small number of studies with better execution and more suitable design or a larger number of studies with less suitable design or weaker execution ( Table2 ). For all designations of strong or sufficient evidence, study results must generally be consistent in direction and size.... In PAGE 5: ... The Task Force makes judgments on the magnitude of effects on a case-by-case basis. Translating Evidence of Effectiveness into Recommendations Effectiveness In general, strength of evidence of effectiveness ( Table2 ) links directly to strength of recommenda- tion (Table 3). Evidence that is inconsistent in direction or size of effect based on definable charac- teristics of the population, setting, or the interven- tion should lead to separate recommendations for different situations.... ..."

### Table 1: Necessary and sufficient validity conditions and (local) change in score for each operator Operator

2002

"... In PAGE 21: ... We use NAY;X to denote the set of nodes that are neighbors of node Y and are adjacent to node X in the current state. The proofs of these results, which are summarized in Table1 , are given in Appendix B. Theorem 15 Let Pc be any completed PDAG, and let Pc0 denote the result of applying an Insert(X;Y;T) operator to Pc.... In PAGE 22: ... There are a number of tricks we can apply to generate more efficiently the candidate operators corresponding to a pair of nodes. Consider the first validity condition for the Insert operator given in Table1 : namely, that the set NAY;X [ T must be a clique. If this test fails for some set T,thenit will also fail for any T0 that contains T.... In PAGE 23: ... In all of the experiments we have performed, however, including those presented in the next section, we have yet to encounter a domain for which GES encounters a state that has too many neighbors. As is evident from the simplicity of the validity conditions from Table1 , there are a number of ways to efficiently update (i.... In PAGE 23: ... Suppose that all the operators have been generated and scored at a given step of (the first phase of) GES, and we want to know whether these operators remain valid and have the same score after applying some operator. From Table1 , we see that if the neighbors of Y have not changed, the first validity condition must still hold for all previously-valid operators; because we are adding edges in this phase, any clique must remain a clique. Furthermore, if the parents of node Y have not changed, we need only check the second validity condition (assuming the first holds) if the score of the operator is higher than the best score seen so far; otherwise, we know that regardless of whether the operator is valid or not, it will not be chosen in the next step.... In PAGE 37: ... Appendix B: Operator Proofs In this appendix, we provide proofs for the main results in Section 5. We show that the conditions given in Table1 are necessary and sufficient for an Insert and Delete operator to be valid for the... In PAGE 39: ... But from Corollary 39, A and C are not adjacent, and thus A ! B C is a v-structure, yielding a contradiction. The conditions from Table1 include checking that some set of neighbors of a node in a com- pleted PDAG are a clique. It follows immediately from Lemma 34 that if any set of neighbors is a clique, then that set of neighbors is a clique of undirected edges.... In PAGE 39: ... B.2 The Insert Operator In this section, we show that the conditions in Table1 are necessary and sufficient for determining whether an Insert operator is valid during the first phase of GES. In particular, we show in Theorem 15 that the conditions hold if and only if we can extract a consistent extension G of the completed PDAG Pc to which adding a single directed edge results in a consistent extension G0 of the com- pleted PDAG Pc0 that results from applying the operator.... In PAGE 41: ... B.3 The Delete Operator In this section, we show that the conditions in Table1 are necessary and sufficient for determin- ing whether a Delete operator is valid during the second phase of GES. In particular, we show in Theorem 17 that the conditions hold if and only if we can extract a consistent extension G of the completed PDAG Pc to which deleting a single directed edge results in a consistent extension G0 of the completed PDAG Pc0 that results from applying the operator.... ..."

Cited by 66

### Table 14. A composite table of significant correlations at the p lt; .01 level.

1998

"... In PAGE 9: ...igure 32. Symptom profile for all conditions combined...............................................47 Table14 .... In PAGE 57: ...01 significance level. Table14 lists the correlations at the a = .01 significance level that show sufficient evidence to reject the null hypothesis (all p lt; .... ..."

Cited by 1

### Table 1 ululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululul SUFFICIENCY

1984

"... In PAGE 31: ...behaviorism and ecologism). Thus, the four meaningful conjunctions of the two positions on these two issues define the four principle modern views shown in Table1 . Of these, Turing-machine functionalism is the strongest and IP the weakest.... In PAGE 31: ... Even if IP turns out to be fatally flawed in one or more ways, being clear about the underlying issues can only help in the ultimate goal of understanding cognition. Thus, the four meaningful conjunctions of the two positions on these two issues define the four principle modern views shown in Table1... ..."

Cited by 6

### Table 1 ululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululululul SUFFICIENCY

1984

"... In PAGE 31: ...behaviorism and ecologism). Thus, the four meaningful conjunctions of the two positions on these two issues define the four principle modern views shown in Table1 . Of these, Turing-machine functionalism is the strongest and IP the weakest.... In PAGE 31: ... Even if IP turns out to be fatally flawed in one or more ways, being clear about the underlying issues can only help in the ultimate goal of understanding cognition. Thus, the four meaningful conjunctions of the two positions on these two issues define the four principle modern views shown in Table1... ..."

Cited by 6