The following changes have been made to HE.RespF_Definitions:Parameters.

1. "Faculty publications" changed to "Faculty research performance" everywhere it appears. Conforms terminology.
2. Added a definition for "Departmental research norm", which is used in a number of research-related functions. The definition is based on a figure to be supplied for the average research per faculty at the database institutions in market segments 1, 2, and 3, adjusted for department and rank/age.
3. In "Departmental academic standing," change "Ave. monthly SPONSORED RESEARCH/departmental research NORM" changed: ____ Ditto for "Faculty Research Performance" under Faculty Database.
4. The variable name "Percent of courses utilizing IT" is changed to "TECHNOLOGY utilization in teaching". Conforms to current usage.
5. "Ave over teaching methods of (course ENROLLMENT/course SECTIONS)/NORMAL CLASS SIZE" changed to "Ave over teaching methods of (course ENROLLMENT/course SECTIONS)–NORMAL CLASS SIZE". Corrects error.
6. "Percent course SECTIONS–TARGET percent course sections]" for Seminars, etc., changed to "Average over course types: ABS[percent course SECTIONS–TARGET percent course sections]" (ABS means Absolute value"). Reduces the number of inputs to Educational Quality by two.
7. Added "Multiplier on the probability for student selection of major" and "Multiplier on the probability for student selection of courses" to the Departmental Performance section. Makes enrollments depend on departmental performance.
8. Replaced "Student Morale" by "Student satisfaction with the department's academic program," which is the average value of student satisfaction: academic for students majoring in the department.
9. Added "Multiplier on maximum class for DL courses and course & major selection for DL students". Will be used in the new Distance Learning procedure.
10. Added "Satisfaction index: assistant professors" and renamed the faculty "Satisfaction index: other than assistant professors." Makes assistant professor satisfaction depend on Player’s input for difficulty of promotion as well as the other factors.
11. Replaced "Prof's Talent Index: TEACHING" SCHOLARSHIP and RESEARCH with "Professor's performance in relation to talent", which is the weighted sum of "performance minus talent" for T, S, and R. Makes faculty satisfaction index depend on performance in relation to talent and the Player-generated institutional priorities, and saves two variables.
12. Added "Strain on discretionary time due to institutional priorities". Makes morale depend negatively on: (a) the deviation between preferred and actual discretionary time; and (b) the degree to which Player has been changing priorities.
13. Added "Academic support ratio" (defined in Section 9.2) to the faculty performance indices for teaching, scholarship, and research. Adjusted the weights accordingly.
14. Specified that the "Prof’s satisfaction indices" in the teaching and scholarship performance functions should be for the prior month, to avoid circularity. These indicators are calculated at the beginning of each trimester based on the teaching data for that trimester. ALSO, specified the research performance indicator to be recalculated monthly to take account of month-by-month changes in research volume.
15. Added "Multiplier on faculty departure probabilities". Makes the transition probabilities for departures depend negatively on faculty satisfaction (morale).
16. Added "Multiplier on promotion probabilities and merit salary adjustments". Makes the transition probabilities for promotions and the overall merit adjustment depend positively on a weighted average of teaching, scholarship, and research performance, with weights based on the Player-generated priority sliders.
17. Added "Sponsored research discretionary time offset percent" in the Faculty Database section. Increments the base preference of faculty for research discretionary time by the indicated percentage of the time devoted to other activities.
18. Added "Policy multiplier on probability of promotion from assistant to associate prof." Connects the newly agreed policy slider to the engine.
19. Added "Probability of accepting an early retirement offer". Depends on the ratio of the offer to annual salary and the faculty member’s overall satisfaction index.
20. Added "Quality drivers for research" in a new section labeled Research Proposals and Awards". The quality drivers are Faculty Research Talent, Faculty Research Performance, Departmental Academic Standing, and Institutional Prestige.
21. Added "Probability of generating a research proposal in a given month" in the same section. Depends on the number of active proposals and projects, the quality drivers for research, and research discretionary time.
22. Added "Size of research proposal: total cost". Depends on departmental parameters and professor's research talent. Note: we decided that all projects will last one year.
23. Added "Probability that a given research proposal will be funded." Depends on the quality drivers for research and the indirect cost rate.
24. Added "Probability that a given research proposal will be funded (evaluated at time of generation)" at the end of the Faculty Database" section.
25. Added "Probability that a student will fail a course" at the beginning of the Student Database section. This implements the academic achievement paragraph in Section 3.1 of Td_2.2. That discussion is amended so that the number of failures is saved in the student sim instead of the academic achievement index.
26. Added "Multiplier on graduation probability for doctoral students". Makes doctoral time to degree depend on the department’s sponsored research and inversely on doctoral teaching load.
27. Variable name "Satisfaction index: academic changed to "Satisfaction index for SL 1-3: academic". Ditto for the student life and athletics variables, except these apply to SL 1 only. Makes clear where these indices apply. "Satisfaction index: academic" values for the other student levels are defined below. "Satisfaction index: student life" and "Satisfaction index: athletics" do not exist for other than traditional undergraduates.
28. Adds "use of IT in teaching" as a driver of Satisfaction index: academic, and changes the weights of the other drivers to accommodate the new driver. Derives from rethinking the IT function.
29. Adds "Satisfaction index for SL-4 (doc): academic". Makes doctoral satisfaction depend on faculty research, etc., as well as course-taking.
30. Adds "Satisfaction index for SL-5 (DL): academic". This applies to the new "Distance learning" student level, which replaces the non-matric level. Deletes library expenditures and increases the weight on "Use of IT in teaching."
31. Added a note under "Satisfaction index: overall" to say that overall satisfaction for SL-2,3, and 5 equals the academic satisfaction index for these students.
32. Added "Satisfaction index: doctoral students" near the end of Student Database.
33. Added weights to the drivers of "Multiplier on percent of alumni who have given anytime during the last five years". Corrects error.
34. Added a new section labeled "Enrollment management functions based on data not available for peer institutions" after "Institution-Level Student Morale Functions". The new section contains the variables "Applications multiplier" and "Yield multiplier", which bring institutional prestige, overall student satisfaction, and enrollment management expenditures into the demand functions for student intake.
35. Added several variables as drivers of "Institutional Prestige" under Trustee Evaluation (now variable 48). Conforms prestige drivers to the Rand study.
36. Added "Growthrate of state appropriations" in a new "miscellaneous" section at the end of the variable list. Makes the growthrate of state appropriations depend on institutional prestige, athletics, and student satisfaction. See section 10 for how the growthrate is used.
37. Added "Effect of urban vs. suburban v. rural location on nontraditional on-campus undergraduate (SL-2) initialization student numbers". Links campus location to the initialization.

1. We decided to add "Institutional Priority" sliders to the interface screen showing the allocation of faculty discretionary time for each department. We ALSO should add sliders for: (a) "Pressure to change teaching load" (–50 to +50 instead of 0-100) under the departmental teaching contact hours report; and (b) "Degree to which priorities are reflected in promotion and salary decisions" (0-100).
2. A new spreadsheet, Faculty Incentives, has been added to the "Response Function" Excel file. It specifies the following functions:
1. Relative institutional priorities. Converts the institutional priority slider values to two sets of relative institutional priorities: (a) relative priority among the full set of faculty activities; and (b) relative priority among teaching, scholarship, and research. The first set is obtained by dividing the slider values by their sum. The second averages the slider values for course preparation, out-of-class student contact, and educational development (the three teaching-related activities), then finds the relative values of the average and the slider values for research and scholarship.
2. Faculty discretionary time. Balances the faculty member’s base discretionary time preferences, adjusted for sponsored research if any, with the allocation implied by the institutional priority. Base discretionary time varies by department and rank as described in Faculty_Templates of HE.GDB.init. The relative importance of the faculty member’s preferences and the institution’s preference are determined by the relation between the relevant faculty talent index and the input slider introduced above, "Degree to which priorities are reflected in promotion and salary decisions. "The discretionary time percentages are updated each trimester subject to latency: suggested lambda = 0.67. This function replaces the one previously specified.
3. Effect of institutional priorities on faculty morale. Calculates the strain due to: (a) deviations between faculty discretionary time preferences and the actual time allocations; plus (b) Player’s "Pressure to change teaching load"; plus (c) Player’s changes in priorities. The resulting figure goes into a response function used in determining of faculty morale.
3. The relative institutional priorities for teaching, scholarship, and research are used in the faculty hiring algorithm as well as for discretionary time determination. This eliminates the need for a separate set of priority inputs on the faculty hiring screen.

1. We decided to eliminate the doctoral admissions target from the interface and make it endogenous on a department by department basis. The target will depend on the department’s sponsored research volume and teaching requirements. For simplicity, let’s assume that doctoral students teach an average of n breakout sections each year (Bo)—early years will be less and later ones more, but we don’t have to track that. Suggest n=1.5.
2. Initialization. (a) For research, calculate docSTUsPerRes$ = initial doctoral student numbers divided by initial direct sponsored research summed over the department’s faculty sims.(b) For teaching, calculate BoSectionsByDocSTUs = MIN[initialBoSections,n*initialDocSTUs], where initialBoSections represents an annual figure..
3. During play, calculate (a) targetDocSTUs = Maximum[docSTUsPerRes$ * sponsoredRes$; initialDocSTUs + BoSectionsByDocSTUs * (currentBoSections – initialBoSections)]: then (b) targetDocIntake = targetDocSTUs –contDocSTUs , where contDocSTUs is "continuing doctoral students" after applying the year-end departure procedure.

1. We decided to add the research overhead rate at the time each proposal is generated to the record for that proposal (and the consequent project if awarded) in the faculty sim. The research overhead rate is determined in the Stage I RA optimization procedure, and changes may be considered, promised, or implemented during the year in the usual way. Implemented changes in the overhead rate are incorporated in new research proposals with immediate effect. However, the lag between proposal and award means that overhead recovery will be phased in over time as in real institutions.
2. Conversions between total research dollars and direct research dollars are as follows:
   total dollars = direct dollars *(1+overhead rate)
   direct dollars = total dollars /(1+overhead rate)
   The propensity of faculty to generate research proposals and the proposal success probability also depend on the overhead rate.
3. We did not discuss the "total dollars" figure associated with individual projects. I think we should include overhead here, and also in the Sponsored Research Office graphs and reports on proposals, awards, and sponsored research volume.
4. We should split the "monthly expenditures on sponsored research expenditures" in the faculty sim into two variables: one that refers to direct expenditures and the other to total expenditures. The effects of research on doctoral requirements, faculty discretionary time, faculty research performance, and any other faculty activity variables are all based on direct research dollars. The financial statements are based on total research dollars, so they now can be calculated by summing the "total research" variable over faculty sims..

1. We decided to include allocations among the Asset Categories shown below. Player can change the asset allocation any time with immediate effect.
2. The parameters (revised from this note’s original draft) are provided on the new HE.RespF_Definitions:Investments spreadsheet and shown below. The expected values have the dimension "percent change per year" net of inflation.



3. The mean and standard deviation for the endowment’s real monthly total return depend on the asset allocation and, following the dividend discount model for investment returns, the change in the inflation rate. Nominal total return equals inflation plus real return. Let f_i be the fraction of the endowment invested in category i. Then:

endowment monthly growthrate =

changeInMonthlyInflationRate +
where w_i is the fraction of the portfolio invested in asset class i
and m_i is the annual expected real return for asset class i.

endowment annual standard deviation =
where the s_ij are the entries in the covariance table.

The formulas for the annual real return and standard deviation are
included in HE.RespF_Definitions:Investments. Applying them
with an asset mix of 50% large-cap stocks, 20% small-cap stocks,
and 30% bonds produces an expected real return of 5.7% and a
portfolio standard deviation of 18.6%.

endowment monthly standard deviation =

Sqrt[endowment annual standard deviation –
12(standard deviation of monthly inflation)²/(1–lambda)]/Sqrt[12]
where lambda is the latency factor for the change in monthly inflation.

4. Endowment[month t] = Endowment[month t–1] *
   (1 + random normal deviate with the above mean and standard deviation +
   any growth or decline due to scenario or chance-card events)
5. Endowment spending is determined at the beginning of each year, based on smoothed market value and the target spending rate. The formula is:
   spending[t] = targetSpendingRate[t]*smoothedMarketValue[t],
   where
   smoothedMarketValue[t] = (1–lambda)*currentMarketValue[t] +
   lambda*smoothedMarketValue[t–1].
   Lambda = 0.65 and currentMarketValue[t] is the endowment’s market value for the last month of the prior year. The "endowment market value" used in Resource Allocation Stage 1 should be set equal to smoothedMarketValue[t].

1. We decided to elaborate distance (DL) learning as a matter of importance especially for the non-research universities. We envision the game’s DL as being mainly web-based or asynchronous rather than of the more facilities-intensive TV-based variety.
2. Student Level 5 will be renamed "Distance Learning Students". CyberCampus will have no non-matriculated students. DL students will use the data in HE.GDB.init:Student_Parameters for non-traditional undergraduates (SL-2) unless otherwise noted. The original data for SL-5, non-matriculated students, can be abandoned.
3. The new SL-5 should be added to the admissions screens and student reports. Player now will set an admissions and track results that same as for non-traditional on-campus undergraduates. Possibly DL students should appear right after on-campus non-traditional students rather than after doctoral students, and we may wish to rename "Non-traditional undergraduates" as "On-campus non-traditional undergraduates."
4. DL students behave like non-traditional undergraduates, except that they don’t place demands on physical facilities or student life budgets. Their enrollments need to be tracked separately and excluded from the determination of normal square footage in the facilities module. SL-5 enrollments already are excluded from the student life response function.
5. The "Lecture without breakout" course category will be renamed "Distance learning courses."
1. The HE.GDB.init:Course_Templates will be expanded to include a column giving the maximum class size for Distance Learning. DL students will take only DL courses. The maximum class size represents a limit on the amount of e-mail, etc., a faculty member can handle during what would have been student contact hours, plus the normal allocation of faculty time for out-of-class student contact.
2. On-campus students (SL 1-4) cannot take DL courses
3. The maximum class size for both DL and on-campus students in all courses will be positive functions of the accumulation of the department’s IT usage level since the beginning of the game.
6. Given the new importance of IT in distance learning, it may be desirable to treat libraries and IT as separate line items in the Stage 2 Resource Allocation model. [Note: we did decide to do this.]
7. The institution’s attractiveness to DL students will be a positive function of the weighted average of the departmental cumulative IT usage scores, with weights equal to the relative preferences of non-traditional undergraduates for choice of major.
8. The maximum class size for DL courses has been added as column E of Dept_Master.
9. Players will input their course mix preferences for conventional courses (seminar, general, lecture with breakout). Departmental faculty have their own intrinsic preferences for course type mix (and also for normal class size). These will be supplied on the HE.GDB.init:Course_Templates. The departmental course preference used each trimester (t) will be calculated as follows:

where x_it^P is the player’s preference for course type i as of time t, x_i^D is the department’s intrinsic preference (which doesn’t vary over time), and l=0.6.

1. The facilities model has been expanded to take account of class type when calculating the effect of enrollments on normal square footage. HE.Facilities now contains separate rows for normal square footage per enrollment and number of enrollments for seminars, general courses and lecture courses with breakouts. (Distance learning courses don’t require space.) Row 32 has been changed to include the new rows and also to correct an error.

1. The inflation rate plays an important role in the financial and resource allocation models, but no specification has been provided so far. Inflation will be on a monthly basis but always reported as an annual rate.
2. initial monthly inflation rate = Max[.01, random normal deviate[.04, .01]]/12
3. monthlyInflationRate[t] =
   monthlyInflationRate[t]+change_in_monthlyInflationRate[t],
   where change_in_monthlyInflationRate[t] =
   (1–lambda)*randomNormalDeviate[0, .005]/12 +
   lambda*change_in_monthlyInflationRate[t–1],
   with lambda = .6. [NOTE: THIS IS SUPERCEEDED BY WFM NOTE OF 19 SEPT.)
4. The monthly inflation rate is used to calculate the return on endowment investments.
5. The annualized inflation rate equals the average of the monthly rates for the prior 12 months. The annualized rate is used in all resource allocation calculations. [SHOULD BE THE SUM OF THE MONTHLY CHANGES: see the WFM Note of 19 September.]
6. The short-term interest rate applied to the operating and capital reserve balances is applied once a year, so it should be based on the annualized inflation rate. (BASE SHORT RATES ARE SUBJECT TO RANDOM VARIATION per EFM Note of 19 Spetember.)
1. Let the annual short-term rate paid on positive balances equal the annualized inflation rate + 0.005 (0.5%).
2. Let the annual short-term rate paid on negative operating reserve balances equal the annualized inflation rate + 0.04 (4%).

1. Non-faculty departmental expense represents the cost of support staff in departments plus expenditures for supplies, travel, etc. It is determined, for the institution as a whole, by Player through stage 2 of the resource allocation optimization.
2. The institution-wide figure should be divided by the institution-wide sum of faculty salaries and direct sponsored research dollars to get the academic support ratio now included in the response functions for faculty performance.
3. There is no need to allocate non-faculty departmental expense to specific departments.

1. We decided to merge "Other operating expense" with "Administration" in the financial statements and in resource allocation Stage 2. The initialization sums for these two quantities should be added together to get the initialization value for the merged quantity.
2. We also decided to have a separate line for "Enrollment management" in the financial statements and resource allocation Stage 2. (This could take the place of "Other operating expense".) This should be initialized at a sum equal to:
   (total student admissions targets at initialization)*(randomNormalDeviate
   with mean 1.0 and standard deviation 0.1)
   (The resulting figure will be in thousands of dollars as desired.) We should ask IRHE to get us a reading on enrollment management expense per admittee.
3. The initial figure for administration should be reduced by the initialization value for enrollment management to avoid changing the initial surplus/deficit.
4. The revenue line "State appropriation" should be added to the financial statements and resource allocation stage I in the case of public institutions only. Initialization values are available and clearly labeled in the games database. The appropriation should grow each year according to his equation:
   appropriation[t] = appropriation[t–1]*(1+inf+respF[t]+randNormDev),
   where "inf" is the annualized inflation rate, respF[t] has been added to the response function set as "Growthrate of state appropriation", and randNormDev has mean 0 and standard deviation 0.0075.
5. We decided to terminate the game if the operating reserve goes negative by a sufficient amount. The limit is based on endowment and plant asset values, but the coefficients are small because of the difficulty of eliminating restrictions on endowment assets and liquidating plant. For now, set the formula to:
   limit[t] = 0.2*smoothedEndowmentMarketValue[t] +
   0.1*replacementValueOfPlant[t]
   The limit should be evaluated and tested at the end of each year.