LOUIS J. CECI, J.
This is a review of a published decision of the court of appeals reversing the judgment of Waukesha County Circuit Court Judge Hafold J. Wol-lenzien in an administrative review action.1 The judgment which the court of appeals reversed was an affirmance of an order of the Wisconsin Employment Relations Commission clarifying the collective bargaining unit for [583]*583negotiations between the Arrowhead school district and the Arrowhead United Teachers Organization. The Wisconsin Employment Relations Commission (hereafter commission) determined that the bargaining unit composed of all full-time and regular part-time “professional employes” of the Arrowhead school district should exclude university students working as intern teachers. The circuit court affirmed the commission’s unit clarification, and the court of appeals reversed the matter, directing that the interns be included in the bargaining unit. We reverse the court of appeals.
On October 5, 1979, the Arrowhead school district (hereafter district) filed a petition for unit clarification, requesting that the commission determine whether the intern teachers employed by the district were municipal employes and whether they should be included in the professional employe bargaining unit. The Arrowhead United Teachers Organization (hereafter union) sought to include the intern teachers within the existing unit. On December 3, 1979, a hearing was held before the commission on the petition, and on June 12, 1980, the commission issued its decision in conjunction with an accompanying memorandum.
The commission found that the teacher interns were “municipal employes” within the meanig of sec. 111.70 (1) (b), Stats., the Municipal Employment Relations Act. The commission also found that the interns were “temporary employees” but not “casual employes” and that the interns did not have a “community of interest” with the full-time and regular part-time professional employes represented by the union. Because of this lack of a community of interest between the interns and the professional employes represented by the union, the commission clarified that pursuant to sec. 111.70(4) (d), the bargaining unit should not include the teacher interns.
The union petitioned the circuit court for review of the commission’s decision, pursuant to secs. 227.15 and 227.-[584]*58416, Stats., on July 11, 1980. The petition asked that the court reverse that portion of the commission’s finding that the teacher interns and professional employes did not share a community of interest. The union concurred with all other findings of the commission. The district challenged the commission’s finding that the interns should be considered temporary rather than casual employes. Had the commission classified the teacher interns as casual employes, they would not qualify as “municipal employes” under the definition of sec. 111.70(1) (b). Therefore, the district requested that the circuit court classify the interns as casual employes, contrary to the commission’s finding.
On July 3, 1981, the circuit court issued its decision, followed by an order and judgment dated July 15, 1981. The court first dealt with the issue of whether the interns constituted municipal employes under sec. 111.70 (l)(b), Stats., and affirmed that portion of the commission’s decision. The judge then addressed the question of whether or not the interns should be included in the existing bargaining unit on the basis of a community of interest and/or the antifragmentation provision as set out in sec. 111.70(4) (d)2.a.2 After comparing the [585]*585job functions, wages, hours, and conditions of employment of both professional employes and teacher interns, the circuit court concluded that the commission had not abused its discretion by finding that no community of interest existed between the two groups of employes. The court then considered the antifragmentation provision found in sec. 111.70(4) (d)2.a., observing that the language is not absolute. The trial judge also stated that in his opinion, the antifragmentation directive is tempered by the language of sec. 111.70(6), which provides that public interest is promoted by allowing municipal employes the opportunity to bargain collectively if they “so desire.” The court then noted that there was no showing that the teacher interns desired to be included in the union’s bargaining unit. Accordingly, the court concluded that the commission had not violated any statutory mandate by its decision and affirmed the commission’s order.
The union then appealed to the court of appeals, asking the court to reverse the circuit court and the commission with respect to their findings excluding the teacher interns from the professional employe bargaining unit. The district did not file any cross-appeal on the issue of the interns’ classification as municipal employes under sec. 111.70(1) (b).
In a lengthy opinion, the court of appeals reversed the circuit court’s judgment based upon a finding that the commission had deviated from its past practice without a satisfactory explanation. The court stated that the commission’s determination of what constitutes a “community of interest” is a conclusion of law and that the [586]*586proper standard of review is found in sec. 227.20(8), Stats.3 Citing Bangor Education Association, WERC Dec. No. 14699 (June 9, 1976'), the court noted that the commission has given “ ‘primary and controlling emphasis to the actual duties and responsibilities of the individual holding the position in question.’ ” 109 Wis. 2d at 377. Because the commission had found the professional employes and teacher interns to have similar duties and responsibilities, the court of appeals concluded that according to its past practice, the commission should have included the two groups within one bargaining unit. The court further noted the commission had also deviated from its past practice concerning unit fragmentation in school districts and had ignored the express statutory limit upon-its discretion as set out in sec. 111.70(4) (d) 2.a. Based upon these observations, the court of appeals concluded that the commission had abused its discretion by these two unexplained deviations and reversed the trial court, directing that the teacher interns be included within the union’s bargaining unit.
Both the commission and the district, as an interested party, have appealed from the decision of the court of appeals. There are two issues which concern us in this review. They are: (1) What is the proper standard for reviewing a commission decision which deviates from prior agency practice and (2) utilizing the proper standard, did the commission abuse its discretion by failing [587]*587to include the teacher interns in the professional employe bargaining unit ?
I.
WHAT IS THE PROPER STANDARD TO BE u UTILIZED IN REVIEWING AN AGENCY DECISION WHICH DEVIATES FROM THE AGENCY’S PRIOR PRACTICE?
Initially, we would like to point out that this appeal involves mixed questions of law and fact. The review encompasses the commission’s construction of sec. 111.70 (4) (d)2.a., Stats., as well as the commission’s application of the statute to the particular set of facts involved in this case. Such issues are questions of law. Bucyrus-Erie Co. v. ILHR Department, 90 Wis. 2d 408, 417, 280 N.W.2d 142 (1979), and Milwaukee v. WERC, 71 Wis. 2d 709, 714, 239 N.W.2d 63 (1976). It also involves purely factual determinations made by the commission, and these determinations are obviously questions of fact. The parties concede that the facts in this case are not disputed.
The union claims, and the court of appeals agreed, that in its past decisions, “the Commission has given primary and controlling emphasis to the actual duties and responsibilities of the individual holding the position in question.” 109 Wis. 2d at 377, citing Bangor Education Association, WERC Dec. No. 14699 (June 9, 1976). See also, Winnebago County (Department of Social Services), WERC Dec. No. 10304-A (September 5, 1979). The court of appeals further stated that the commission has never found that individuals performing similar duties under similar working conditions did not share a “community of interest.” Because the commission found that the teacher interns perform similar duties under [588]*588similar working conditions as do the regular teachers, yet refused to find that both groups shared a “community of interest,” the court of appeals concluded that the commission had deviated from its prior practice.
When we review past commission decisions, it does indeed appear that in the majority of the decisions, the commission has found the disputed individual to share a community of interest with the bargaining unit from which the individual has sought to be separated. Wisconsin Heights School District, WERC Dec. No. 17182 (August 7, 1979) ; Madison Metropolitan School District, WERC Dec. No. 14161-A (January 7, 1977) ; Kenosha Unified School District No. 1, WERC Dec. No. 18431 (March 11, 1975).4 We then must refer to the,statutory standards for reviewing agency decisions, as contained in § 227.20, Stats., entitled “Scope of review.”5 Section 227.20(8) provides:
“The court shall reverse or remand the case to the agency if it finds that the agency’s exercise of discretion is outside the range of discretion delegated to the agency by law; is inconsistent with an agency rule, an officially stated agency policy or a prior agency practice, if deviation therefrom is not explained to the satisfaction of the court by the agency; or is otherwise in violation of a constitutional or statutory provision; but the court shall not substitute its judgment for that of the agency on an issue of discretion.”
[589]*589Thus, the plain language of sec. 227.20(8) states that exercises of discretion which are inconsistent with the agency’s prior practice are not fatal if “deviation therefrom” is satisfactorily explained to the court. In order to determine what amounts to a satisfactory explanation, we must direct our attention to prior case law.
This court noted that sec. 227.20(8) provides an unusual standard of review in Wis. Asso. of Manufacturers & Commerce v. PSC, 100 Wis. 2d 300, 305, 301 N.W.2d 247 (1981). In Public Serv. Corp. v. Public Serv. Comm., 109 Wis. 2d 256, 325 N.W.2d 867 (1982), this court also dealt with the review of an agency’s decision under sec. 227.20 (8). The court stated in that opinion:
“While this court has recognized that inconsistencies in determinations arising by comparison are not proof of arbitrariness or capriciousness, if the deviation is so unreasonable as to be without a rational basis or the result of an unconsidered, willful and irrational choice of conduct, the decision is an arbitrary and capricious one. Robertson Transport Co. v. Public Serv. Comm., 39 Wis. 2d 653, 159 N.W.2d 636 (1968) ; Olson v. Rothwell, 28 Wis. 2d 233, 137 N.W.2d 86 (1965). Agencies are not free to shift between decisions without a rational basis.” Public Serv. Corp. v. Public Serv. Comm., 109 Wis. 2d at 263 (emphasis added).
Thus, it appears that a satisfactory explanation of a deviation from prior agency practice under sec. 227.20 (8) must be grounded upon a rational basis. We, therefore, hold that the proper standard of review under sec. 227.20(8), Stats., is that deviations from past agency practice will be sustained if grounded upon a rational basis.
The court of appeals also found that the commission had deviated from its prior practice regarding the anti-fragmentation provision found in sec. 111.70(4) (d)2.a. [590]*590Although the commission decided shortly after its order in this case that separate bargaining units were appropriate despite this provision, City of Franklin, WERC Dec. No. 18208 (November 4, 1980), once again a review of the decisions reveals that in the past, this factor has led to the inclusion of the disputed position within a single unit. Germantown School District, WERC Dec. No. 17494 (December 11, 1979) ; Gateway VTAE District, WERC Dec. No. 17449 (November 19, 1979); Maple School District, WERC Dec. No. 17463 (November 28, 1979). Therefore, the same standard concerning the rational basis test also applies to this deviation from past practice concerning fragmentation of bargaining units.
II.
DID THE COMMISSION EXERCISE PROPER DISCRETION BY EXCLUDING THE TEACHER INTERNS FROM THE PROFESSIONAL EMPLOYES’ BARGAINING UNIT?
As we have previously stated, past commission decisions usually involve a finding of a shared “community of interest” where individuals have similar responsibilities and working conditions, are involved in the same integrated task force, etc. Kenosha Unified School Dist. No. 1, WERC Dec. No. 13431 (March 11, 1975). To decide whether or not the commission has utilized a rational basis in its departure from past decisions concerning the “community of interest,” we first must determine how the term is utilized by the agency. However, the term “community of interest” is used nowhere in sec. 111.70(4) (d), which deals with the commission’s powers. Instead, it appears to be an accepted concept of [591]*591labor law and is often utilized in the area of collective bargaining. As one treatise has characterized the term:
“The importance of this concept of a community of interests may be attributed primarily to the fact that the designated unit and its chosen representative must operate for the mutual benefit of all employees. Should a dissimilarity of interests exist, then surely collective bargaining would be severely impeded. ... A single agreement for a bargaining únit composed of employees with diverse interests would lead to unrest, chaos, and possibly disruption of production.” J. Abodeely, R. Hammer, and A. Sandler, The NLRB and the Appropriate Bargaining Unit, p. 12 (Rev. ed. 1981) (emphasis added).
The court of appeals concluded that a shared community of interest arises automatically from similar duties and working conditions. However, when this court reviews prior commission decisions that deal with the concept of community of interest, two factors come to our attention. First, we are not aware of any past decision which dealt with professionals and interns who have virtually identical duties, as do the regular teachers and the teacher interns. Thus, the instant case appears to have been a question of first impression for the commission. Secondly, it appears that the commission’s decisions to group individuals within a single bargaining unit have not been based solely upon the shared “community of interest” which automatically arises from similar duties and working conditions, as the court of appeals believed. Rather, the commission has consistently considered the following factors in deciding whether individual employes are to be grouped within a single bargaining unit:
“1. Whether the employes in the unit sought share a ‘community of interest’ distinct from that of other employes.
“2. The duties and skills of employes in the unit sought as compared with the duties and skills of other employes.
[592]*592“3. The similarity of wages, hours, and working conditions of employes in the unit sought as compared to wages, hours, and working conditions of other employes.
“4. Whether the employes in the unit sought have separate or common supervision with all other employes.
“5. Whether the employes in the unit sought have a common work place with the employes in said desired unit or whether they share a work place with other employes.
“6. Whether the unit sought will result in undue fragmentation of bargaining units.
“7. Bargaining history.” City of Franklin, WERC Dec. No. 18208 (November 4, 1980) ; Wisconsin Heights School District, WERC Dec. No. 17182 (August 7, 1979) ; Kenosha Unified School District No. 1, WERC Dec. No. 13431 (March 11, 1975).
We believe that the court of appeals viewed the commission’s criteria too simplistically when it stated that similar duties and working conditions automatically lead to a shared “community of interest.”6
The commission has never articulated precisely what constitutes a shared community of interest among employes. However, when reviewing the commission’s decisions, it appears that the concept involves similar interests among employes who also participate in a shared purpose through their employment. In Brown County, WERC Dec. No. 1231 (January 8, 1974), the commission determined that research analysts should be grouped within the same bargaining unit as the social workers [593]*593and psychologists. The commission noted in reaching this decision that:
“The research analyst performs research functions and develops statistics to assess the effectiveness of the programs and techniques used by the social workers and psychologists . . . the Research Analyst works in support of the same program engaged in by the Social Workers and Psychologists.” Id.
Thus, it appears that the commission utilizes the concept to embody an interest shared by employes whose work involves a common purpose. Therefore, by combining these individuals within a single bargaining unit, collective bargaining is not undermined because of restlessness arising from widely dissimilar interests.
We previously stated that the construction of a statute is a question of law and that this court is not bound by an interpretation utilized by an agency. Berns v. Wis. Employment Relations Comm., 99 Wis. 2d 252, 261, 299 Wis. 2d 248 (1980). Pursuant to sec. 227.20(5), this court may always review questions of law.7 Bucyrus-Erie Co. v. ILHR Department, 90 Wis. 2d at 417. The concept of community of interest arises out of the commission’s construction of sec. 111.70(4) (d)2.a. of the Municipal Employment Relations Act. We have also stated that “ ‘. . . the construction and interpretation of a statute adopted by the administrative agency charged by the legislature with the duty of applying it is entitled to great weight’ ” and that it is “ ‘. . . only when the interpretation by the administrative agency is an irrational one that a reviewing court does not defer to it.’ ” [594]*594Beloit Education Asso. v. WERC, 73 Wis. 2d 43, 67, 242 N.W.2d 231 (1976). We have qualified this standard, though, by stating that where the question involved is one of first impression, the court will accord the agency’s interpretation due weight in determining the appropriate statutory construction, rather than the great weight-rational interpretation standard. Berns v. Wis. Employment Relations Comm., 99 Wis. 2d at 261. The great weight-rational interpretation standard applies in situations where the “commission’s interpretation reflects a practice or position ‘long continued, substantially uniform and without challenge by governmental authorities and courts.’ ” Id., citing Wood County v. Bd. of Vocational, T. & A. Ed., 60 Wis. 2d 606, 618, 211 N.W.2d 617 (1973).
Although the question of determining whether a shared community of interest exists between students and professionals who have almost identical duties and working conditions is a question of first impression for the commission, we believe that the general determination of an appropriate bargaining unit involves the application of MERA and requires the commission’s expertise. We find that the application of the community of interest concept does indeed reflect a long-standing practice of the commission in determining appropriate bargaining units under MERA. Wisconsin Heights School District, WERC Dec. No. 17182 (August 7, 1979); Madison Metropolitan School District, WERC Dec. No. 14161-A (January 7, 1977); Kenosha Unified School District No. 1, WERC Dec. No. 13431 (March 11, 1975). Section 111.70(4) (d)2.a., Stats., was created by ch. 124, Laws of 1971, effective November, 1971, and has apparently been applied consistently by the commission. The record does not indicate previous challenges to the application of this concept. Blackhawk Teachers’ Federation v. [595]*595WERC, 109 Wis. 2d 415, 423, 326 N.W.2d 247 (Ct. App. 1982). Therefore, we hold that the interpretation given to sec. 111.70(4) (d)2.a., through the commission’s development and application of the community of interest concept, is entitled to the great weight-rational basis standard.
Invoking this deferential standard of review, we further hold that the commission’s use of this concept involves a proper interpretation of sec. 111.70(4) (d)2.a. The statute provides that the commission shall determine the appropriate bargaining unit. In determining the appropriate unit, the statute states the commission may decide whether employes in “the same or several . . . professions or other occupational groupings constitute a unit.” The statute does not mandate that employes with similar duties must be grouped within a single unit. The commission, therefore, utilizes a concept which embodies similar interests shared by employes working toward a common purpose in order to decide whether they should constitute a single unit. We believe that the commission’s use of the concept of community of interest constitutes a rational basis for the interpretation of sec. 111.70 (4) (d)2.a., which charges the commission with the duty of determining appropriate bargaining units.
After finding that the commission’s use of the concept involves a proper interpretation of sec. 111.70(4) (d)2.a., we then must determine whether or not the commission’s departure from its prior practice utilizing this concept is sustained by a rational basis. When we review the record in the instant case, pursuant to sec. 227.20(1), Stats.,8 we find that the commission relied on the follow[596]*596ing facts in making its decision.9 The commission found that each intern is employed by the district for only one semester, receiving a stipend of $2,000 during this term of employment. The stipend’s value is in reality reduced by the fact that the interns are registered as students and must pay tuition at their respective colleges for the credits they receive during their internship. The only benefits which the interns receive are five days’ sick leave.
Students for the internship position are selected by the Wisconsin Improvement Program, a program administered by the University of Wisconsin. An internship is one of two ways in which the students may fulfill the requirement of student teaching in order to obtain their teaching credentials. Although the district interviews the intern applicants, they seldom reject any students.10 Six or seven interns are employed each semester by the district. The district assumes the right to discipline the interns should any such need arise, although none had at the time this case was commenced. The interns receive no formal training, but attend the same orientation and “in-service” as beginning regular teachers. Each has a cooperating teacher who is supposed to function as a “mentor” and evaluates the intern’s performance for his or her college. Each intern also has supervising instructors from the college who are supposed to visit the district. The actual time spent by interns with their cooperating teachers and supervising instructors appears to vary widely and to have no consistent pattern.
[597]*597The interns are not certified as regular teachers and possess “intern teaching licenses.” These licenses do not restrict their work, and the commission noted little difference between their responsibilities and those of the regular teachers. Interns are assigned to fewer classes than regular teachers and to more supervisory periods, such as study halls, etc. Therefore, they are able to spend more time during the school day in preparation of classes. Interns who teach art and physical education classes are allowed the option of spending time at elementary schools, so that they may be certified for grades kindergarten through 12. Regular teachers are not allowed this option. Interns also receive time off to interview for employment in other school districts.
The record indicates that the district has hired eighty-nine full- and part-time regular teachers since it began utilizing teacher interns in the 1973-74 school year, only seven of which had served as interns. The testimony also shows that the school administrator’s testimony is that the district’s primary purpose for participating in the internship program is to use the interns to teach its students, while the educational benefits incurred by the interns are secondary to the district.
The commission concluded that interns are employed by the district for the primary purpose of teaching students in much the same manner as regular teachers and function as such. However, the commission concluded from the district’s track record in hiring students to fill positions as regular teachers that a low expectation of continued employment existed among the interns. Accordingly, the commission concluded that the two groups shared no community of interest.
The trial court noted that in Madison Metropolitan School District, WERC Dec. No. 14161-A (January 7, 1977), the commission found a shared community of interest based upon similar “job functions, wages, hours and conditions of employment.” After comparing these [598]*598objective factors for interns and regular teachers, the trial court concluded that the commission had not abused its discretion by finding no community of interest based upon dissimilarities between the two groups.
We agree with the trial court. We also believe that these dissimilarities, when considered in conjunction with the following factor noted by the commission, provide the rational basis for the commission’s departure from its prior practice of finding a shared community of interest:
“While they perform similar work under similar conditions, the aspirations of the teachers, as a group, are presumably directed toward a career-length employment; but the interns, regardless of the purposes for which the District hires them, can be expected to have interests centering on their opportunities for learning, training, practice and eventual hire elsewhere, and their concerns in collective bargaining would logically focus on elemental and short-term subjects.” (Emphasis added.)
The unique interests of the interns centering around the educational opportunities available to them through the internship program, namely, the opportunity to obtain their teaching credentials, clearly prevent them from sharing a community of interest with the regular teachers. 11 In spite of the district’s reason for employing in[599]*599terns, the interns remain primarily students. Therefore, they are necessarily concerned with the short-term benefits which the internship offers, while the regular teachers are concerned with the long-term benefits of career-length employment. The very fact that the interns are concerned with long-term employment outside the Arrowhead school district prevents them from working toward the same purpose as the regular teachers. We also believe that the internship is directly related to their educational program, as opposed to economic interests, as is indicated by the low stipend they receive, as well as the fact that they are supplied by a program coordinated by the University of Wisconsin. They also remain registered as students during the internship. It is our view that academic interests are alien to the usual employment relationship, which centers around economic interests. We believe that these academic interests would not fall easily within the collective bargaining process, which would be concerned mainly with career-length objectives such as benefits, wages, retirement, etc. Therefore, we find the required satisfactory explanation pursuant to sec. 227.20(8), Stats., for the commission’s deviation from its prior practice concerning the interns’ unique interests. This is in keeping with sec. 227.20 (10), which, along with logic, dictates that the commission’s experience in matters concerning the determination of appropriate bargaining units be given due weight.12
[600]*600The union attacks this deviation as focusing on the “subjective” characteristics of the interns’ positions, as opposed to “objective” characteristics such as wages and conditions of employment, which the union claims has been the commission’s past focus. However, it is clear to us that this concentration on the subjective characteristics of the disputed individual’s position has been considered by the commission in the past. In Madison Joint School District, WERC Dec. No. 14814-A (December 7, 1976), the commission determined that clerical and secretarial employes had a community of interest separate and distinct from that of custodial and maintenance employes. In reaching this conclusion, the commission stated that in determining the appropriate bargaining unit under sec. 111.70(4) (d)2.a., it was aware of “the need for ensuring that the unique interests of a given group of employes will not be subordinated to the interests of another bargaining unit.” The commission further stated that because of this need, the commission looks to the facts of a particular case to determine what is the appropriate bargaining unit. The commission has held consistently to this view and, as recently as May, 1982, stated the following:
“This Commission has interpreted Section 111.70(4) (d)2.a. to mean that at times there is a need for a mix of bargaining units which afford employes the opportunity to be represented in workable units by organizations of their own choosing, which may reasonably be expected to be concerned with the unique interests and aspirations of the employes in said units.” Madison Water Utilities Employees Association, WERC Dec. No. 19584, p. 10 (May 10, 1982; emphasis added).
The National Labor Relations Board operates under a provision somewhat similar to Wisconsin’s MERA.13 [601]*601The NLRB has dealt with the issue of whether graduate students teaching as research and teaching assistants at their colleges share a community of interest with the full-time faculty. The board found that the students’ employment depended on their status as graduate students and that the. students received no faculty benefits. Although the fact situation involved is somewhat different from the case at hand because the Arrowhead interns receive five days’ sick leave and are not graduate students teaching at their colleges, we believe that the facts are similar enough to merit our consideration. In that instance, the NLRB concluded that although the students carried out some of the same duties as did the faculty, they remained primarily students and, therefore, did not share a sufficient community of interest to lead to their inclusion within a single unit. Adelphi University, 195 N.L.R.B. 639, 640 (1972). See also, St. Clare’s Hospital and Health Center, 229 N.L.R.B. 1000 (1977). The NLRB has made similar decisions concerning psychology student interns working at health care institutions and law student interns providing legal services to low-income persons. The Clark County Mental Health Center, 225 N.L.R.B. 780 (1976), and Evergreen Legal Services, 246 N.L.R.B. 964 (1979).
The court of appeals also found that the commission had deviated from its prior practice concerning fragmentation of bargaining units. As we have stated, when we [602]*602review past decisions, it appears that the commission has been inclined toward avoiding separate bargaining units. Germantown School District, WERC Dec. No. 17494 (December 11, 1979) ; Gateway VTAE District, WERC Dec. No. 17449 (November 19, 1979) ; Maple School District, WERC Dec. No. 17463 (November 28, 1979). This is due to the antifragmentation provision of sec. 111.70 (4) (d)2.a., Stats., which provides that the commission “shall whenever possible avoid fragmentation by maintaining as few units as practicable in keeping with the size of the total municipal work force.” The commission has interpreted this provision as follows:
“This statutorily mandated antifragmentation policy does not in all cases necessarily preclude the creation of separate units. Rather, by providing that this policy is to be followed ‘whenever possible,’ the statute implicitly acknowledges that separate units can be established under certain circumstances.”
Kenosha Unified School District No. 1, WERC Dec. No. 13431 (March 11, 1975).
As we stated above, this statutory provision became effective in November of 1971. The court of appeals and the union have made it clear that it has been applied consistently by the commission, and the record does not indicate that it has been challenged by the government or the courts. Blackhawk Teachers’ Federation v. WERC, 109 Wis. 2d at 423. Therefore, the great weight-rational basis standard also applies to this provision.
Utilizing this standard, we must agree with the commission’s interpretation of the provision. The plain language of sec. 111.70(4) (d)2.a. indicates that the commission “shall whenever possible” avoid fragmentation. It does not mandate inclusion of employes within a single unit in all circumstances, and the commission has not acted in the past as if it has. Rather, we agree with the [603]*603commission that the statute “implicitly acknowledges” that separate units are appropriate in special situations.
We also believe that the same reasons given for the finding of a lack of community of interest between the regular teachers and intern teachers provide a rational basis for the commission’s departure from its past practice concerning fragmentation. The interns are primarily students, concerned with the academic, rather than economic, benefits of their employment by the district. Their interests have been characterized as “short-term” by the commission. If they are grouped within the single unit consisting of the regular teachers, their special interests would clearly be in the minority and may potentially go unrecognized. We, therefore, find that the commission has provided a rational basis for its deviation concerning fragmentation of bargaining units.
It must be kept in mind that under sec. 111.70(4) (d), the legislature has given the commission discretion to determine the appropriate units for collective bargaining. This court has recognized that the exercise of discretion by an agency involves more than simply making a choice between alternatives without stating the reasons behind the choice. Reidinger v. Optometry Examining Board, 81 Wis. 2d 292, 297-98, 260 N.W.2d 270 (1977). We further stated:
“This process must depend on facts that are of record or that are reasonably derived by inference from the record and a conclusion based on a logical rationale founded upon proper legal standards. As we pointed out in State v. Hutnik (1968), 39 Wis. 2d 754, 764, 159 N.W.2d 733, ‘. . . there should be evidence in the record that discretion was in fact exercised and the basis of that exercise of discretion should be set forth.” Id., citing McCleary v. State, 49 Wis. 2d 263, 277, 182 N.W.2d 512 (1971).
[604]*604In this case, we find that there is evidence that the commission did in fact exercise its discretion as shown by the facts set forth in its memorandum.
In summary, we hold that the commission properly exercised its discretion by excluding the teacher interns from the professional employes’ bargaining unit. We also hold that the commission satisfactorily explained its deviation from prior practice under sec. 227.20(8), Stats., by finding that the interns remained primarily students, with the corresponding interests of students, notwithstanding the fact that their duties and responsibilities were similar to those of regular teachers. The decision of the court of appeals reversing the circuit court is reversed, and the circuit court’s opinion is affirmed.
By the Court. — The decision of the court of appeals is reversed; the judgment of the circuit court is affirmed.