1 Rosenberg, Maurice, Civil Justice Research and Civil Justice Reform, 15 Law & Soc'y Rev. 473 (19801981); Sarat, Austin, The Role of Courts and the Logic of Court Reform: Notes on the Justice Department's Approach to Improving Justice, 64 Judicature 300 (1981).CrossRefGoogle Scholar

2 . At about the same time, the American Bar Association launched its Action Commission to examine the cost (and delay) problem, and to propose solutions. See Janofsky, Leonard S., A.B.A. Attacks Delay and the High Cost of Litigation, 65 A.B.A.J. 1323 (1979); Hufstedler, Seth & Nejelski, Paul, A.B.A. Action Commission Challenges Litigation Cost and Delay, 66 A.R.A.J. 965 (1980); Paul Nejelski, With Justice Affordable for All, Judges' J., Summer 1980, at 4; Leonard S. Janofsky, Facing the Crisis of Court Costs and Delay, Bar Leader, Jan. 1982, at 22–24, 35.Google Scholar

3 . Trubek, David M., Sarat, Austin, Felstiner, William L. F., Kritzer, Herbert M., & Grossman, Joel B., The Costs of Ordinary Litigation, 31 U.C.L.A. L. Rev. 72 (1983) (hereinafter cited as Trubek, Ordinary Litigation). That article is in fact an extended summary of our final report, David M. Trubek et al., Civil Litigation Research Project Final Report (Madison: University of Wisconsin Law School, 1983) (hereinafter cited as Trubek, CLRP Report), copies of which are available at cost from the University of Wisconsin Law School. The analysis presented here is a detailed presentation of one aspect of the analysis described in “Ordinary Litigation”; the only significant change is in the analytic model used to account for lawyer time. The results for the revised model are, with minor exceptions, the same as for the earlier analysis; the differences that do appear in the two sets of results do not lead to any differences in overall interpretations or conclusions. Reports of additional findings from CLRP can be found in Richard E. Miller & Austin Sarat, Grievances, Claims, and Disputes: Assessing the Adversary Culture, 15 Law & Soc'y Rev. 525 (1980–81); Joel B. Grossman et al., Measuring the Pace of Litigation in Federal and State Trial Courts, 65 Judicature 86 (1981) (hereinafter cited as Pace of Litigation); Grossman et al., Dimensions of Institutional Participation: Who Uses the Courts and How? 44 J. Pol. 86 (1982); Herbert M. Kritzer, The Judge's Role in Pretrial Case Processing: Assessing the Need for Change, 66 Judicature 28 (1982); Herbert M. Kritzer & Jill K. Anderson, The Arbitration Alternative: A Comparative Analysis of Case Processing Time, Disposition Mode, and Cost in the American Arbitration Association and the Courts, 8 Just. Sys. J. 6 (1983); Herbert M. Kritzer, The Civil Litigation Research Project: Lessons for Studying the Civil Justice System, in Proceedings of the Law and Justice Statistics Work (1984); Herbert M. Kritzer et al., Courts and Litigation Investment: Why Do Lawyers Spend More Time on Federal Cases? 9 Just. Sys. J. 7 (1984) (hereinafter cited as Court Effect); Herbert M. Kritzer et al., The Impact of Fee Arrangement on Lawyer Effort, Law & Soc'y Rev. (forthcoming) (hereinafter cited as Fee Effect); Herbert M. Kritzer et al., Lawyers and Litigation (forthcoming).Google Scholar

4 One could think of these costs as the “opportunity cost of money and time”.Google Scholar

5 Though in fact we ask judges to evaluate such “costs” all the time in areas like child custody, e.g., in determining what is in the best interest of the child.Google Scholar

6 . See Kritzer, Herbert M., Fee Arrangements and Fee Shifting: Lessons from the Experience in Ontario, 47 Law & Contemp. Probs. 125 (1984).Google Scholar

7 See James S. Kakalik & Abby E. Robyn, Costs of the Civil Justice System: Court Expenditures for Processing Tort Cases (Santa Monica, Cal.: Rand Corporation, 1982); J. S. Kakalik & R. L. Ross, Costs of the Civil Justice System: Court Expenditures for Various Types of Civil Cases (Santa Monica, Cal.: Rand Corporation, 1983). For CLRP's effort in this area, see Terence Dungworth, The Institutional Cost of Civil Disputes, in 3 Trubek, CLRP Report, supra note 3.Google Scholar

8 We use the term disputant to refer to litigants (as opposed) to their lawyers; we use the term participant to encompass both litigants (disputants) and their lawyers.Google Scholar

9 This is referred to as “indirect” costs by Earl Johnson, Jr., et al., Access to Justice in the United States: The Economic Barriers and Some Promising Solutions, in M. Cappelletti & B. Garth, Access to Justice: A World Survey (Milan: Guiffre, 1978). Johnson, id. at 919, divides out-of-pocket expenses (which he refers to as “direct” costs) into five subcategories: (1) official charges, (2) security for costs, (3) auxiliary costs, (4) special bonds, and (5) lawyer's fees.Google Scholar

10 We define the “typical” case in terms of the median; in this example, the median case is that for which half the cases devoted a greater proportion of their expenditures to lawyer's fees and half devote a smaller proportion.Google Scholar

11 The second quartile corresponds to the median (or middle) case; the first quartile shows the cut-off for the bottom quarter (25%) of cases, and the third quartile shows the cut-off for the top quarter. Including the information for the first and third quartiles gives a picture of how much the proportion (in this case) varies. If the first and third quartiles are close to the median (as is true for individuals in fig. 2, A), then most cases are fairly similar; if the quartiles diverge sharply from the median (which is true for organizations in fig. 2, B), then there is a lot of variation in the cases on the characteristics of interest.Google Scholar

12 The value of the disputant's time was estimated by multiplying the amount of time spent, as reported by the disputant, by an hourly rate. Where possible, we used as a hourly rate the respondent's wage rate. If the wage rate was unknown (or nonexistent if the respondent was not employed), we used a figure equal to 150% of the minimum wage; when we tried out a variety of figures (minimum wage and various multiples of the minimum wage), we found that the specific figure chosen had virtually no impact on our results because few disputants reported spending a substantial block of time.Google Scholar

13 Only 31% of the lawyers we spoke to reported that any other lawyers in their firm had worked on the case.Google Scholar

14 There are potentially many problems involved in trying to account for success, not the least of which is whether retrospective data on stakes can be used in the calculation of the ratio; in one sense it is possible that a low success ratio, which would be computed (for plaintiff) as recovery divided by stakes, simply reflects a poor calculation of stakes on the part of the lawyer (or client). We have examined the success question elsewhere. See Trubek, Ordinary Litigation, supra note 3, at 109, and Kritzer et al., Lawyers and Litigation, supra note 3, ch. 7.Google Scholar

15 . See Grady, John F., Some Ethical Questions About Percentage Fees, 2 Litigation 20–26, 52–53 (1976).Google Scholar

16 Of the 1,382 lawyers involved in court cases who responded to our extended interview, 56 were paid on a flat fee basis, 757 on an hourly fee basis, 431 on a contingent fee basis, and 138 on some other basis (i.e., they were employees of corporations, legal aid, prepaid legal service plans, or their fee was to be paid on some combination of hourly, flat, and/or contingent).Google Scholar

17 For the median case, expenses constitute 8% of the total bill.Google Scholar

18 For cases involving $100,000 or less (89% of cases for which we have stakes information), the regression coefficient is. 80 (r2 = .13); for cases involving $10,000 or less (52% of cases), the regression coefficient is 2.54 (r2 = .05).Google Scholar

19 One question this image suggests, which we will not at present try to answer, is what brings an end to the action-reaction process?.Google Scholar

20 . The analysis of initiatives would be greatly complicated by the two-way linkage shown in fig. 7. It would also require a type of data set that is almost impossible to obtain; see Herbert M. Kritzer, Studying Disputes: Learning from the CLRP Experience, 15 Law & Soc'y Rev. 503, 520 (19801981).Google Scholar

21 We have examined this model for contingent-fee lawyers elsewhere. See 2 Trubek, CLRP Report, supra note 2, at II-117; Kritzer et al., Lawyers and Litigation, supra note 2, ch. 6; and Kritzer et al., Fee Effect, supra note 3.Google Scholar

22 One might expect that this would simply reflect differences in the kinds of lawyers and law firms used by corporations and individuals; while this is substantially true, corporate law firms sometimes do serve individuals and often bill individuals at a lower rate than organizations.Google Scholar

23 The twelve courts were: Urban: Federal District Court, Western Pennsylvania. Federal District Court, Central California. Philadelphia Court of Common Pleas. Los Angeles County Superior Court (Downtown Branch). Midurban: Federal District Court, Eastern Wisconsin. Federal District Court, New Mexico District Court for the Second Judicial District, Albuquerque. Circuit Court, Milwaukee, Wisconsin. Chester County Court of Common Pleas (Pennsylvania). Rural: Federal District Court, South Carolina. Richland County Court of Common Pleas (South Carolina). Dodge County Circuit Court (Wisconsin).Google Scholar

24 For the other cases handled by these lawyers, we conducted “mini-interviews,” which provided very limited data; 430rnini-interviews were completed. In addition to the court case respondents, we completed another 188 long interviews and 69 mini-interviews with lawyers involved in cases handled by alternative institutions. A small number of additional interviews were completed with lawyers who handled cases that did not go to courts or alternatives. In all, 2,082 lawyer interviews were completed.Google Scholar

25 Because of lack of money, we had to drop a randomly selected subsample of 59 cases from the interviewing process.Google Scholar

26 A total of 167 cases were screened out because the case appeared to be an uncontested collections case or an uncontested divorce or the case was disposed of in a fashion that indicated it may well have been refiled somewhere else.Google Scholar

27 We could not locate her, she refused to be interviewed, or she indicated that her involvement in the case was so limited that she would have little or nothing to tell us about the case. A comparison of cases for which we did and did not attempt to complete interviews can be found in 1 Trubek, CLRP Report, supra note 2, at 1–120. The cases not included in the interviewing process tended to be simpler (as measured by pretrial activity) and were disposed of more quickly.Google Scholar

28 Id.Google Scholar

29 See Trubek, Ordinary Litigation, supra note 3.Google Scholar

30 See 2 Trubek, CLRP Report, supra note 3, at 11–100.Google Scholar

31 We did not have data on all 29 variables for every case in the regression subset. Since we could not eliminate all cases with missing data and have an adequate sample for analysis, where possible we estimated these missing items, using means or medians. To avoid artificially depressing the variance in our data, we added a normally distributed random number to each of the means or medians that replaced missing data.Google Scholar

32 The breakdown between plaintiffs and defendants for our analysis of hours and rates is: Plaintiffs Defendants Total Hours…123 278 401 Rates…106 240 346.Google Scholar

33 . For more detail on the data collection design, see I Trubek, CLRP Report, supra note 2, vol. 1; Kritzer, Herbert M., Miller, Richard E., & Felstiner, William L. F., Studying Disputes by Survey, 25 Am. Behav. Sci. 67 (1981).CrossRefGoogle Scholar

34 We use numbers in brackets to “index” each variable in the tables that follow in the analysis section.Google Scholar

35 Discovery-related motions were classified as discovery events; consequently, motions refers only to nondiscovery motions.Google Scholar

36 Elsewhere we haw used the term pace; Grossman et al., Pace of Litigation, supra note 3.Google Scholar

37 We included in duration time taken by appeals (if those appeals were evident in the court record);.Google Scholar

38 . Galanter, Marc, Why the “Haves” Come Out Ahead: Speculations on the Limits of Legal Change, 9 Law & Soc'y Rev. 95 (1974).Google Scholar

39 This measure is taken from John P. Robinson & Phillip R. Shaver, Measures of Social Psychological Attitudes 102 (Ann Arbor, Mich.: Institute for Social Research, 1964).Google Scholar

40 . Earl Johnson, Jr., Lawyer's Choice: A Theoretical Appraisal of Litigation Investment Decision, 15 Law & Soc'y Rev. 567 (198081).Google Scholar

41 The technique of analysis used to examine the impact of the variables summarized in tables 1 and 4 on hours and rates was linear regression analysis. Before applying the method, we had to deal with three other methodological problems: multicollinearity, outliers, and heteroscedasticity. We will briefly summarize our concerns in each of these areas here; a more complete treatment can be found in 2 Trubek, CLRP Report, supra note 3, at 11–105. We examined all our predictor variables for possible multicollinearity (high intercorrelations) problems. This appeared to create problems only for our original standard operating procedure indicators. TheSOP variables used in the analysis, pretrial events SOP [25] and estimating case value SOP [26], were factor scores created in order to alleviate this problem (we started with four SOP indicators). One problem with regression analysis is that it is greatly affected by “outliers” (i.e., cases with extreme values on one or more key variables). Our data set was such that we expected to have outliers since our random sampling procedure did pick up the occasional “big case.” To examine the impact of outliers on our results, we performed our regression analysis with and without the outliers in the data set. For the hours regression, the inclusion or exclusion of the outliers had minimal impact on our results, and consequently we include them in the results reported below. For the rate regression, three outlying cases, with hourly rates in excess of $200 per hour, were removed from the analysis because their inclusion greatly altered the results. The last problem, heteroscedasticity, only applied to the hours regression. Regression analysis requires the assumption that the variance of the equation “error term” (i.e., the difference between the observed and predicted values of the dependent variable) be constant for all systematically identifiable subsets of observations. When this assumption is not met, one has the problem of heteroscedasticity; the impact of this problem is to severely weaken the significance tests that are ultimately carried out. Common sense suggests that one should expect a greater range of predictive error for big cases than for small cases; this is consistent with the notion that stakes serve to “cap” the level of investment of time (and money) in a case. To adjust for the heteroscedasticity in our data, we applied an adjustment factor to each observation; this yielded a set of “corrected” regression equations. The specific adjustment used was to divide all of the variables for each case by the square root of stakes. For most analytic purposes, the “corrected” results were employed, and it is these results that we, by and large, report below. This adjustment was not needed for our analysis of hourly rates.Google Scholar

42 Since our goal in assessing the overall fit is to look at the “best” fit we can get, we have used the R2 from the uncorrected (“ordinary least squares”) regression estimates; because of the mathematics of the technique, this estimate of the equation will always yield a better fit for the equation than will the “corrected” (“weighted least squares”) regression estimates. On the other hand, the corrected estimates yield the best information regarding the significance of the contribution of individual variables or groups of variables. It is worth noting that the one major impact of excluding 30 cases (out of the total of 401) that have extreme values on one or more of the variables in the equation increases the R' to. 49 (the adjusted R2 increases from. 34 to .45).Google Scholar

43 Our work on outcomes fails to reveal any significant improvement in results from the client's perspective as the level of lawyer expertise increases; Kritzer et al., Lawyers and Litigation, supra note 3, ch. 7.Google Scholar

44 See Johnson, supra note 40.Google Scholar

45 This analysis is reported in Kritzer et al., Court Effect, supra note 3.Google Scholar

46 Cf. the discussion in Trubek, Ordinary Litigation, supra note 3.Google Scholar

47 See note 22 supra for the classification of courts into the three categories.Google Scholar

48 See Janofsky, Crisis of Court Costs and Delay, supra note 1.Google Scholar

49 . Galluccio, Nick, The Rise of the Company Lawyer, Forbes Mag., Sept. 18, 1978, at 168; Stanley A. Kaplan, Some Ruminations on the Role of Counsel for a Corporation, 56 Notre Dame Law. 873 (1980); Doug Lavine, Corporate Legal Units Moving Up, Nat'l L.J., Oct. 9, 1978, at 1.Google Scholar

50 If one assumes about a 50% increase in rates charged by lawyers since 1978 (the year to which the data apply), the typical hourly rate now is probably something like $75.Google Scholar

51 As we suggested above, we do not have a good explanation for the basis of the finding.Google Scholar

52 See 2 Trubek, CLRP Report, supra note 3, at 11–58.Google Scholar

53 A recent study of rules to control discovery abuse found relatively little abuse to control; see Joy A. Chapper & Roger A. Hanson, Cost-shifting in Maricopa County Superior Court: An Examination of Rule 37(a)(4), 8 Just. Sys. J. 325 (1983).Google Scholar

54 2 Trubek, CLRP Report, supra note 3, at 11–58.Google Scholar

55 . See Wasby, Stephen L., The Function and Importance of Appellate Oral Argument: Some Views of Lawyers and Federal Judges, 65 Judicature 340 (1982); Woodford Howard, J. Jr., Query: Are Heavy Caseloads Changing the Nature of Appellate Justice 66 Judicature 57 (1982).Google Scholar

56 See Marvin E. Frankel, Partisan Justice (New York: Hill & Wang, 1978); William H. Simon, The Ideology of Advocacy: Procedural Justice and Professional Ethics, 1978 Wis. L. Rev. 29.Google Scholar

57 . Brazil, Wayne D., The Adversary Character of Civil Discovery: A Critique and Proposals for Change, 31 Vand. L. Rev. 1295 (1978).Google Scholar