Class Action Sex Discrimination Prevention Controls of AIS FACU= LTY PROMOTION- A FORENSIC AUDIT, EXPERT WITNESS TESTIMONY & COMPUTER LITIGA= TION SUPPORT CASE

&nb= sp;

Avi Rushinek, &

Sara Rushinek

University= of Miami

Correspond= with Sara Rushinek, arush@miami.edu=

*** =

*** Some of the tables have been omitt= ed to save time and space please contact the authors in writing for more informat= ion.

I<= span style=3D'mso-tab-count:1'> Abstr= act

= ;

= ;

The authors = develop a methodology to control for “Class Action Sex Discrimination” = law suits using AIS FACULTY PROMO= TION as a case in point. This is a FORENSIC AUDIT, EXPERT WITNESS TESTIMONY & COMPUTER LITIGATION SUPPORT = CASE study approach to minimize the monetary damage, flag emerging cases and providing quick Settlement of a Class that is pending. Such systems can be used by accreditation organizations such as the AACSB, who accredits accounting programs and examined how their publication rate. Academic organizations such as the= AAA (American Accounting Association) can use it to promote equal opportunities= and compliance with the EEOC (Equal Employment Opportunity Commission). The EEOC can use it, to enforce compliance, and it can be used by universities, as well as law firms.<= /o:p>

<= /o:p>

<= /o:p>

Using standa= rd statistical methods, the authors found that universities promote AIS (Accounting Information Systems) faculty primarily on the basis of publicat= ions as they were statistically significant.&nb= sp; In contrast, the “gender” or Sex were statistically insignificant, but indicate a bias in favor of female faculty members in the sample. That means that there= is no evidence of sex discrimination in the population yet, but the sample sugges= ts that universities may be biased against males and in favor of female faculty members in promotion decisions. That maybe a red flag that eventually may develop to a class sex dis= crimination case.

<= /o:p>

II &nb= sp; I INTRODUCTION

The proliferation of “Class Action Sex Discrimin= ation” cases reported by the press raises the issue of Prevention Controls. This study looks at the case of AIS FACULTY PROMOTION from a FORENSIC AUDIT, EXPERT WITNESS TESTIMONY & COM= PUTER LITIGATION SUPPORT point of view. We will review some of the more prominent cases that the press discu= ssed more recently.

Such cases include but are not limited to the followin= g: http://writ.news.= findlaw.com/sebok/20040809.html reports about “The Huge Class Action Sex Discrimination Suit against Wal-Mart:
Should It Proceed as a Class Action, or be Decertified?” By ANTHONY J. SEBOK, anthony.sebok@brooklaw.edu on Monday,= Aug. 09, 2004. “Earlier this summer, = a federal district court in San Francisco certified a sex discrimination class action suit against Wal-Mart. (Whe= n a court "certifies" a class action, it allows the case to proceed w= ith a class of plaintiffs, according to class action rules - as opposed to with= a number of individual plaintiffs.)

The class is claiming= that Wal-Mart discriminated against women in the way it recruited and promoted managers. The issue of whether this is true has not yet been determined. Wal-Mart has claimed that its experts will show that there is no evidence of discrimination against women in stores, and the court noted that Wal-Mart w= ill have a chance to prove this at trial.”

The class web site, http://genderclassactionagains= tcostco.com/costco94.pl reports another similar case “Are you a current or former female employee of Costco who sou= ght advancement to management?…..About the Costco Lawsuit

A nationwide employment discrimination class = action lawsuit has been filed against Costco Wholesale Corporation. The suit charg= es that Costco operates a "glass ceiling" at the store-management le= vel which precludes women from obtaining promotion to assistant manager and gen= eral manager positions. On Thursday, January 11, 2007, Judge Marilyn H. Patel certified the class finding that three current and former Costco employees = may represent all women employed by Costco in the United States denied promotio= n to assistant and/or general manager positions since January 3, 2002. “

http://www.genderl= awsuitagainstsmithbarney.com/ reports about another similar case= of “Gender Discrimination Lawsuit Against Smith Barney Plaintiffs alleged that Smith Barney discriminated against female Financial Advisors in compen= sation and business opportunities. The parties entered into a four-year settlement agreement that is subject to Court approval.”

This study uses the number = of refereed publications as a measure of “research productivity.”<= span style=3D'mso-spacerun:yes'> It develops a model to forec= ast the faculty rank (1 assistant, 2 associate and 3 full professor ) as a func= tion of the number of publications and the faculty gender (Sex=3D1 for males, an= d 0 for females).

The purpose of this study w= as to develop a regression analysis connected to a continuous online survey that = will raise red flags for the existence of sex discrimination. This is a tool for compliance fore= nsic audit with regulations of government agencies, such as the EEOC (Equal Employment Opportunities Commission) in the US. The regression helps determine whe= ther sex is a significant factor in promotion decisions. If it is significant, then it may indicate that a class action may form at a future time, and enable the part= ies to reduce its likelihood before it is too late.

We are using data collected= by Simkin et. al. but applying it in a different way, and suggesting another data collection methodology for the future.&nbs= p; Unlike Simkin et. al who focus on the “profile of the publicat= ion rates of accounting faculty, and gender” for academic purposes, we are looking at it more from the forensic audit, expert witness testimony, and litigation support point of view.

Thus our goal is to create benchmarks that might be useful to decide whether gender or sex discriminat= ion exists on a continuous basis in an automated cost effective method of an internet WWW server online survey methodology. For that purpose we recommend usin= g Microsoft SharePoint software that= lets a lay person such as an administrator that is not an IT (Information Techno= logy) professional to design and implement online interactive continuous survey, = fast and effectively.

III &n= bsp; REVIEW OF THE LITERATURE OF RESEARCH PRODUCTIVITY

&nbs= p; We focus on providing expert witness testimony and computer litigation support= for employment and labor law discrimination cases, in general and class actions= in particular. There are many ot= her reasons why measures of research productivity are important to accounting faculty. As Simkin et. al. su= ggest that it is important “in creating benchmarks for the quantity and qua= lity of publications needed to obtain tenure, be promoted, or receive annual mer= it awards (Hasselback, et. al, 2000; Hult, et. al., 1997, Pierce and Garner, 1= 995; Hagerman and Hagerman, 1989) Then too, administrators have a similar intere= st in these norms .. (Maranto and Streuly, 1994; Campbell, et. al., 1983; Schu= ltz, et. al., 1989).

&nbs= p; Faculty research productivity is also important to those universities and colleges seeking accreditation (Bublitz and Kie, 1984; Hasselback, 2000). One application in the accounting = area is for comparisons among similar institutions (Lucertini, 1995; Posy and Pa= rker, 1989; Jacobs, et. al., 1986), comparisons between Ph.D. and non-Ph.D. grant= ing institutions (Schultz, et. al, 1989), or studies seeking to identify the accounting profession’s most prolific researchers (Hech and Bremsyn, 1986; Hech, et. al, 1991; Gopolakrishan and Chandy, 1991).” Some research correlates publicati= ons to the area of specialization. (Campbell and Morgan ,1987; Englebrecht, et. al.,1994).

Gender studies of research productivity have been common, but none of them focu= s on sex discrimination and compliance testing.= As Simkin et. al. suggest “Cole and Zuckerman (1984) note, for example, that over 50 such studies of academics outside the accounting discipline have found that fema= le faculty members publish significantly fewer articles than their male counterparts. In a later study, Zuckerman (1987) found that women publish between 50 and 60 percent as many papers of what men publish.” However, for the purpose of class = action employment litigation is it significant?&n= bsp;

&nbs= p; Gender differences apply to the accounting areas as Streuly and Maranto [1994] found th= at men published slightly more articles, but that these differences were not statistically significant. Dw= yer [1994] confirmed the finding academics (per Hasselback’s Accounting Faculty Directory) that women had published fewer times in academic journals, but is it statistically significant and up-to-date?

= As schools increase the publications requirement for promotion (Read, Rama, and Raghunandan, 1998; Zivney, 1995; Milne and Vent, 1987, Englebrecht, et. al, 1994), testing for sex discrimination may be more important. This study foc= uses on “premier” journals articles (Van Fleet, et. al., 2000, Hasselback, et. al., 2000). However, it is possible that the other journal may have the same behavior patterns (Hasselback, et. al., 2000; Heck, et. al, 1990 and 1991).

The quality of a journal is= a subjective matter and varies among researchers (see, for example, Zivney, e= t. al., 1995). AIS faculty member often publish in computer publications, which may not be considered as Acco= unting Journals. As Daigle and Arnol= d (2000) found that the “top 25 AIS researchers” listed non-accounting publication outlets as the top journals.&n= bsp; Likewise, Christiansen, et. al., (2002) suggests that the tendency applies to accounting Ph.D.’s in that study’s researcher cohort= .

Adjusting for co-authors (Hasselback, et. al., 2000) or accounting specialty (Campbell and Morgan, 1987), may further confound the results (Alexander and Mabry, 1994; Brown a= nd Gardner, 1985). “Leveling the playing field” among the research facilities of organizations may further confound the finding (Englebrecht, = et. al, 2000, Read, et. al., 1998, Dwyer, 1994, Schultz, et. al., 1989

IV &n= bsp; = ; MICROSOFT SHAREPOINT ONLINE SURVEY- ADVANTAGES OF AN ONLINE COMPARED TO A TRADITIONAL SURVEY

&nbs= p; As Simkin et. al. note “Many of the studies cited above are now at least= ten years old or used statistical methods that some scholars suggest hide more = than they show.” The trouble= with paper surveys is that they are so expensive, and time consuming. Traditional survey also involves m= any un-needed intermediaries that slow down the process.= Such middle men as chairperson of the accounting department at each school are “requesti= ng that this individual in turn distribute them to the faculty members in their departments,” which is yet another middle man in the process of data collection. Furthermore, such surveys are typically based on printed directories, such as “faculty = and school listings in Prentice Hall’s Accounting Faculty Directory 2001-= 2002 (Hasselback’s).,” which may be outdated as soon as they are published. In contrast, schoo= ls update their web sites much more frequently than any printed publication. Therefore, using an online survey = that picks the information from the internet WWW can be more up-to-date then a printed directory.

We have used the data that Simkin et. al. collected as follows: “The areas of research specialization were (1) accounting information systems (AIS), (2) auditing, (Audit), (3) financial accounting (Financial), (4) managerial accounting (Management), (5) tax accounting (Ta= x), and (6) other (Other). The Other category captured all responses that did n= ot list the first five areas above, and included governmental and not-for-prof= it accounting, international accounting, accounting education, and ethics. Our= academic rank categories were (1) Adjunct Professor, (2) Assistant Professor, (3) Associate Professor, (4) Full Professor, and (5) Other.”

We have focused on AIS (Accounting Information Systems) faculty member’s ranks, Assistant, Associate, and Full professors, th= eir number of publications, and whether they were males (1) or females (0). This way we had a much more homoge= neous data, which would result in a higher R-Square in a multivariate linear regression analysis.

V &nbs= p; DATA ANALYSIS AND RESULTS:

We entered the information from the Simkin et. al. (= 2007) into Microsoft SharePoint server survey (Mark G. Simkin Stephen Kerr Rich= ard Mason, "PUBLICATION RATES OF ACCOUNTING FACULTY:A MULTIDIMENSIONAL PROFILE" 2007 Oxford Business & Economic Conference, ISBN : 978-0-9742114-7-3) demonstrating how organizations can automate and continuously audit the results. We have further exported the results from Microsoft SharePoint ser= ver survey into excel for some additional analysis. As reported by a Google search &= #8220;Publication Rates Of Accounting Faculty: A Multidimensional Profile PUBLICATION RATES. OF ACCOUNTING FACULTY<= /b>: A MULTIDIMENSIONAL PROFILE. Mark G. Simkin. Stephen Kerr= . Richard Mason. trintrin.com/obec/Mark%20G.%20Simkin,%20Stephen%20Kerr,%20Richard%20M= ason.doc

The Publication Rates Of Accounting Faculty: A Multidimensional Profile PUBLICATION RATES. OF ACCOUNTING FACULTY:. A MULTIDIMENSIONAL PROFILE” divided the “survey instruments into a computerized data set. Auditing of the entered data resulted in an approxim= ate error rate of .2% for the 15% of the data subjected to random review. The t= wo data-entry errors found were corrected.” Their data “provides mean publication rates for all faculty combined, separated by area of research specialization and gender. These results reflect the reported number of peer-reviewed publications divided by the number of respondents reporting t= hat specialization. Thus, the first panel of Table 1 provides the combined resu= lts for all 202 of the tenured or tenure-track faculty (i.e. assistant, associa= te, and full professors) “

Table A

SUMMARY OUTPUT

Table A

AIS Rank=3Df(Pubs,Sex)

Regression Statistics

Multiple R

0.971985=

R Square

0.944755=

Adjusted R Square

0.907925=

Standard Error

0.271403=

Observations

6<= /span>

The SUMMARY OUTPUT in Table A describes the AIS rank a= s a function of Pubs (Publications) = and Sex (Male or Female), AIS Rank=3Df(Pubs,Sex).

&nbs= p; &= nbsp; &nbs= p; &= nbsp;

Regression Statistics include the Multiple R 0.971985, R Square 0.944755, = and Adjusted R Square &n= bsp; 0.907925, which shows that the Publications and Sex explain about 90% of the variatio= ns in rank promotion decisions for AIS faculty. Likewise, the table shows the Stan= dard Error of 0.271403, and the Observations 6. &nbs= p; &= nbsp;

The Regression Statistics show the Multiple R with a v= alue of 0.971985, and the R Square with a value of 0.944755, as well as the Adjusted R Square with a value of 0.907925, or about 90%. It shows that a combination of Publications and Sex explains about 90% of the variation in Rank.

Following is a description of the faculty rank values = and their position descriptions:

Rank Legend = ;

Rank = Description of Rank

1 &nb= sp; Assistant Professors

2 &nb= sp; Associate Professors

3 &nb= sp; Full Professors

Table B

ANOVA		Table B	AIS Rank=3Df(Pubs,Sex)
	df	SS	MS	F	Significance F
Regression	2<= /span>	3.779021=	1.88951<= /o:p>	25.65186=	0.012985=
Residual	3<= /span>	0.220979=	0.07366<= /o:p>
Total	5<= /span>	4<= /span>

	Coefficients	Standard Error	t Stat	P-value	Lower 95%	Upper 95%	Lower 95.0%	Upper 95.0%
Intercept	1.034536<= /o:p>	0.206693<= /o:p>	5.005184<= /o:p>	0.015349*=	0.376747<= /o:p>	1.6923249=	0.376746971	1.69232486
Pubs	0.085692<= /o:p>	0.011964<= /o:p>	7.162661<= /o:p>	0.005605*=	0.047618<= /o:p>	0.123766<= /o:p>	0.047618181	0.123765976
Sex	-0.359907=	0.227225<= /o:p>	-1.58392<= /o:p>	0.211378**	-1.083039=	0.3632256=	-1.08303909	0.363225632

* Significant at t= he .o5 level of statistical significance

** Not significant at the .o5 le= vel of statistical significance

The ANOVA (Analysis of Variance) in Table B support the summary in Table A, it shows the statistical significance of the model of A= IS Rank=3Df(Pubs,Sex). Table B i= ncludes the following parameters: df (Degrees of Freedom), SS (Sum of Squares), MS (Means Squers), and the F value and the Significance F. For the Regression, these paramete= rs have the following values of: 2, 3.779020662, 1.889510331, 25.65185978, and 0.012984867, respectively= .

Most importantly, the last 2 value reflect the test of hypothesis, rejecting the null hypothesis that entire model have zero correlation with rank with an F-value of 25.65185978 and a statistical leve= l of confidence of 0.012984867, which is better than at the .05 level of statist= ical significance.

In contrast with the F and the Significance F for the entire model, only some of the individual variables are significant. The first 2 namely the intercept and the Pubs are significant, but t= he Sex is not significant at the .o5 level.&n= bsp; The 2^nd part of this ANOVA test shows the Coefficients, Standard Error, t Stat, and P-value, for&n= bsp; the Intercept 1.034536, 0.206693, 5.005184, 0.015349 and the Pubs 0.085692, 0.011964, 7.162661, 0.005605 respectively.

Mo= st important are the pairs of t Stat and P-value for each of the variables. The P-values show the statistical significance being lower than the .o5 required to reject the null hypothesi= s, saying that in the population these variables are greater than zero, and therefore, statistically significant in making rank promotional decisions.<= span style=3D'mso-spacerun:yes'> It suggests that the number of publications is a significant factor in promoting assistant to associate, a= nd associate to full professor of AIS.

However, for the last variable Sex, the opposite is true, it is insignificant at the= .o5 level. The Sex, a dummy variable (1 for male a= nd 0 for female) has the following information after the Pubs in the table, it s= hows the: Coefficients, Standard Error, t Stat, and P-value: Sex=3D-0.359907, 0.227225, -1.58392, and 0.211378>.o5, which is the = most important. The P-value shows = that Sex is not statistically significant at the .o5 level, since it is higher t= han .o5. Therefore, we cannot rej= ect the null hypothesis that is saying that in the population there is basicall= y no Sex discrimination, which a coefficient that is equal to zero.

Another important piece of information is the Coeffici= ents of Publications and Sex. These coefficients reveal that on the average the Pubs with a value of 0.08569207= 9, implies that it takes 1 publication per rank. Or, in other words, as Assistant profes