In re Election of the United States Representative for the Second Congressional District
This text of 653 A.2d 79 (In re Election of the United States Representative for the Second Congressional District) is published on Counsel Stack Legal Research, covering Supreme Court of Connecticut primary law. Counsel Stack provides free access to over 12 million legal documents including statutes, case law, regulations, and constitutions.
Opinions
Borden, J.
This is an action brought pursuant to General Statutes § 9-3231 by Edward W. Munster and [605]*605Sam Gejdenson, in separate petitions, each challenging the results of the election canvass and recanvass (election) for the office of the United States Representative for the Second Congressional District of Connecticut (district).2 Munster was the candidate of the Republican Party, and Gejdenson was the candidate of [606]*606the Democratic Party. Accordingly, their names appeared on all ballots prepared for and used in the election.
The election was held on November 8,1994.3 Because, the results of the canvass of returns from the election [607]*607were sufficiently close, a mandatory recanvass was conducted pursuant to General Statutes § 9-311a.4 Following that recanvass, the secretary of the state declared [608]*608that 186,030 votes had been cast, that Gejdenson had received 79,160 votes and that Munster had received 79,156 votes. According to those figures, therefore, Gejdenson would have been declared the winner by a margin of four votes.
I
Procedural History
On November 18, 1994, both Munster and Gejden-son filed, pursuant to § 9-323,5 essentially simultaneous petitions with Ellen A. Peters, the chief justice of the Supreme Court. This is the first time in its history that § 9-323 has been invoked. Pursuant to that statute, on November 21, 1994, Judge Aaron Ment, the chief court administrator, designated Justice David M. Borden and Justice Robert I. Berdon as the two additional members of the panel to adjudicate the claims [609]*609made in the respective petitions.6 Also pursuant to that statute, notice of the petitions, and of the hearing to. be held thereon to begin on November 29, 1994, was given by personal or abode service to Munster, Gejden-son, the secretary of the state, the elections enforcement commission, David Bingham and Howard E. Proper Lamchick; see footnote 2; the town clerks of the fifty-four towns comprising the district, and the moderators of the election proceedings within the district. In all, approximately 200 notices were served.7
The hearings began on November 29,1994, pursuant to the order of notice. At the outset, the panel entered several procedural orders. We granted the motion of the secretary of the state and the elections enforcement commission to be made parties to the proceeding. See footnote 2. Upon motions by Munster, we entered two orders. First, we ordered the town clerks, who are the official custodians of all election materials, until further order of the panel, to preserve all election materials used in the election, including actual ballots, absentee ballot materials, such as envelopes, and moderators’ reports.8 Second, we ordered that the secretary of the state, the treasurer and the comptroller not take any [610]*610action pursuant to General Statutes § 9-315,9 until our further order, regarding the election, including the making of any count and the declaration of any person as having been elected as the United States Representative from the district.
Before the start of the evidentiary hearing, we also considered whether this proceeding would be governed by the usual rules of evidence applicable to a trial. The parties took different positions on the issue. Munster argued that because this proceeding was sui generis, the rules of evidence should not apply and that we should instead apply the rules of admissibility governing administrative proceedings under the Uniform Administrative Procedure Act. See General Statutes § 4-178.10 Gejdenson argued that the rules of evidence [611]*611should apply. We ruled that, although this is a sui generis proceeding insofar as it requires Justices of the Supreme Court not only to determine the law but to find the facts, it is nonetheless also a judicial, rather than an administrative, fact-finding proceeding and that, therefore, we would apply the rules of evidence, exercising our discretion in favor of admissibility insofar as possible. We conducted this proceeding in accordance with this ruling.
For five full days of trial, beginning on November 30, 1994, and ending on December 7, 1994, we heard evidence. Despite the constraints imposed by the time limitations required by § 9-323, 11 each party had a full and fair opportunity to present his evidence and claims. In accordance with the schedule that we established, [612]*612Munster filed his posttrial brief on December 9, 1994, Gejdenson filed his posttrial brief on December 12, 1994, and we heard final oral argument on the afternoon of December 12, 1994.
II
Substantive Claims
Before turning to the substantive claims of the parties, it is useful to set forth what this case does not involve. Neither party has made any claim of criminal conduct. Neither party has made any claim of fraudulent conduct. Neither party has made any claim of willful, deliberate or intentional misconduct. None of the evidence presented by the parties would, in any manner, have supported any such claims. We turn, therefore, to the specific claims that the parties have advanced regarding the election and the recanvass held in the district.
A
THE NORWICH RECOUNT
Both Munster and Gejdensen have claimed that the vote count and recanvass in the town of Norwich were inaccurate as a result of the use of new voting technology in that town. Both candidates were credited with fewer votes in the recanvass than in the original vote count on election night, and, therefore, they both requested that we order a manual recount of the ballots in that town.
Pursuant to No. 94-225 of the 1994 Public Acts,12 the town of Norwich was designated as one of three [613]*613towns13 for a demonstration project for the use of electronic equipment for the casting and counting of ballots in the November, 1994 election. Accordingly, the voting in Norwich for this election was conducted pursuant to the “1994 Marksense Demonstration Project” (demonstration project), rather than pursuant to the usual combination of mechanical machine voting and mechanical tabulation of votes for voters who voted at the polls, and paper ballots and hand counting of the votes of those who voted by absentee ballot, a combination used uniformly in this state for many decades.
Under the demonstration project, a voter who voted at the polls was given a paper ballot on which were indicated the candidates for all of the offices involved in the election. Approximately 3/4 of one inch to the right of each candidate’s name was, in light black ink, an [614]*614abbreviation of the candidate’s party endorsement, namely, “Rep” for Munster and “Dem” for Gejden-son. Immediately to the right of the party designation was, in heavy black ink, the depiction of what may be described as a horizontal interrupted arrow approximately 9/16 of one inch in total length, the interrupted or blank portion of which was approximately 3/16 of one inch in length.
Free access — add to your briefcase to read the full text and ask questions with AI
Borden, J.
This is an action brought pursuant to General Statutes § 9-3231 by Edward W. Munster and [605]*605Sam Gejdenson, in separate petitions, each challenging the results of the election canvass and recanvass (election) for the office of the United States Representative for the Second Congressional District of Connecticut (district).2 Munster was the candidate of the Republican Party, and Gejdenson was the candidate of [606]*606the Democratic Party. Accordingly, their names appeared on all ballots prepared for and used in the election.
The election was held on November 8,1994.3 Because, the results of the canvass of returns from the election [607]*607were sufficiently close, a mandatory recanvass was conducted pursuant to General Statutes § 9-311a.4 Following that recanvass, the secretary of the state declared [608]*608that 186,030 votes had been cast, that Gejdenson had received 79,160 votes and that Munster had received 79,156 votes. According to those figures, therefore, Gejdenson would have been declared the winner by a margin of four votes.
I
Procedural History
On November 18, 1994, both Munster and Gejden-son filed, pursuant to § 9-323,5 essentially simultaneous petitions with Ellen A. Peters, the chief justice of the Supreme Court. This is the first time in its history that § 9-323 has been invoked. Pursuant to that statute, on November 21, 1994, Judge Aaron Ment, the chief court administrator, designated Justice David M. Borden and Justice Robert I. Berdon as the two additional members of the panel to adjudicate the claims [609]*609made in the respective petitions.6 Also pursuant to that statute, notice of the petitions, and of the hearing to. be held thereon to begin on November 29, 1994, was given by personal or abode service to Munster, Gejden-son, the secretary of the state, the elections enforcement commission, David Bingham and Howard E. Proper Lamchick; see footnote 2; the town clerks of the fifty-four towns comprising the district, and the moderators of the election proceedings within the district. In all, approximately 200 notices were served.7
The hearings began on November 29,1994, pursuant to the order of notice. At the outset, the panel entered several procedural orders. We granted the motion of the secretary of the state and the elections enforcement commission to be made parties to the proceeding. See footnote 2. Upon motions by Munster, we entered two orders. First, we ordered the town clerks, who are the official custodians of all election materials, until further order of the panel, to preserve all election materials used in the election, including actual ballots, absentee ballot materials, such as envelopes, and moderators’ reports.8 Second, we ordered that the secretary of the state, the treasurer and the comptroller not take any [610]*610action pursuant to General Statutes § 9-315,9 until our further order, regarding the election, including the making of any count and the declaration of any person as having been elected as the United States Representative from the district.
Before the start of the evidentiary hearing, we also considered whether this proceeding would be governed by the usual rules of evidence applicable to a trial. The parties took different positions on the issue. Munster argued that because this proceeding was sui generis, the rules of evidence should not apply and that we should instead apply the rules of admissibility governing administrative proceedings under the Uniform Administrative Procedure Act. See General Statutes § 4-178.10 Gejdenson argued that the rules of evidence [611]*611should apply. We ruled that, although this is a sui generis proceeding insofar as it requires Justices of the Supreme Court not only to determine the law but to find the facts, it is nonetheless also a judicial, rather than an administrative, fact-finding proceeding and that, therefore, we would apply the rules of evidence, exercising our discretion in favor of admissibility insofar as possible. We conducted this proceeding in accordance with this ruling.
For five full days of trial, beginning on November 30, 1994, and ending on December 7, 1994, we heard evidence. Despite the constraints imposed by the time limitations required by § 9-323, 11 each party had a full and fair opportunity to present his evidence and claims. In accordance with the schedule that we established, [612]*612Munster filed his posttrial brief on December 9, 1994, Gejdenson filed his posttrial brief on December 12, 1994, and we heard final oral argument on the afternoon of December 12, 1994.
II
Substantive Claims
Before turning to the substantive claims of the parties, it is useful to set forth what this case does not involve. Neither party has made any claim of criminal conduct. Neither party has made any claim of fraudulent conduct. Neither party has made any claim of willful, deliberate or intentional misconduct. None of the evidence presented by the parties would, in any manner, have supported any such claims. We turn, therefore, to the specific claims that the parties have advanced regarding the election and the recanvass held in the district.
A
THE NORWICH RECOUNT
Both Munster and Gejdensen have claimed that the vote count and recanvass in the town of Norwich were inaccurate as a result of the use of new voting technology in that town. Both candidates were credited with fewer votes in the recanvass than in the original vote count on election night, and, therefore, they both requested that we order a manual recount of the ballots in that town.
Pursuant to No. 94-225 of the 1994 Public Acts,12 the town of Norwich was designated as one of three [613]*613towns13 for a demonstration project for the use of electronic equipment for the casting and counting of ballots in the November, 1994 election. Accordingly, the voting in Norwich for this election was conducted pursuant to the “1994 Marksense Demonstration Project” (demonstration project), rather than pursuant to the usual combination of mechanical machine voting and mechanical tabulation of votes for voters who voted at the polls, and paper ballots and hand counting of the votes of those who voted by absentee ballot, a combination used uniformly in this state for many decades.
Under the demonstration project, a voter who voted at the polls was given a paper ballot on which were indicated the candidates for all of the offices involved in the election. Approximately 3/4 of one inch to the right of each candidate’s name was, in light black ink, an [614]*614abbreviation of the candidate’s party endorsement, namely, “Rep” for Munster and “Dem” for Gejden-son. Immediately to the right of the party designation was, in heavy black ink, the depiction of what may be described as a horizontal interrupted arrow approximately 9/16 of one inch in total length, the interrupted or blank portion of which was approximately 3/16 of one inch in length. The point of the arrow aimed at the candidate’s party designation and, further to the left, his name. At the top of each ballot was the following instruction, which we have reproduced in reduced format as it appeared on the ballot:
To vote, complete the arrow at the RIGHT of the candidate’s name of your choice. To vote for a candidate whose name is not printed on the ballot, write the name of the candidate in the line provided and complete the anow. __
The voting took place as follows. After having been properly identified and checked, the voter was given such a paper ballot. The moderator’s handbook, issued for the demonstration project by the secretary of the state, contains the following suggestions: “When issuing a ballot, the ballot clerk should offer the voter a ‘privacy sleeve,’ which is a plastic or cardboard sleeve into which the ballot can be inserted so that the markings on the ballot can not be seen. Although each elector should be offered a ‘privacy sleeve’ he is not required to take it. It is also recommended that when the Ballot Clerk issues a ballot to an elector, the Clerk should say something like ‘Vote for just one per office, and connect the front and back of the arrow.’ ”
The voter then entered a voting booth where he or she marked his ballot with either a number 2 pencil or a pen supplied by the manufacturer of the “marksense machine,” a machine designed to read and tabulate the ballots inserted into it. The voter then proceeded to the [615]*615machine and inserted the ballot into it for counting. The voter then was required immediately to leave the polling place.14
The absentee ballots were in essentially the same form as those used at the polls. After having been properly processed and delivered to the appropriate polling places, the absentee ballots were inserted into the marksense machine for reading and tabulation.
After the mandatory recanvass ordered by the secretary of the state following the November 8 election, the recanvass moderator for Norwich certified to the secretary of the state that Munster had received 3802 votes and Gejdenson had received 5711 votes. Neither Munster nor Gejdenson, however, had confidence in the accuracy of these figures. Accordingly, they filed the petitions in this case.
In that part of his petition devoted to Norwich, Munster alleged that “sufficient inaccuracies and irregularities occurred in the original count and recanvass to require the handcounting of all ballots.” He cited an example of an unaccounted discrepancy of ten votes between the number of votes registered by the “memory pack”15 of a particular machine on election day and the number of votes registered by that memory pack during the recanvass. Munster claimed that this called “into question the validity of the entire count on that machine and the reliability of all the machines.” With respect to the absentee ballots, Munster claimed that the technology “showed a similar inconsistency between Election Day and the Recanvass,” citing a one vote discrepancy between the number of absentee bal[616]*616lots registered by the machines on election day and during the recanvass. “Furthermore,” Munster alleged, “the overall count for the candidates on the Business Record machines between Election Day and the Recan-vass changed by thirteen votes. While such a variation is perhaps understandable when humans do the counting, there is no similar rationale for a supposedly infallible machine count. Accordingly, all the ballots cast in Norwich should be recounted by hand . . . .”
Gejdenson’s petition was confined to the Norwich count. He described the system used as follows: it “uses paper ballots on which the voter is to fill in a line between two points to indicate his/her selection. A special scanning device ‘reads’ the mark and records it accordingly. The advantage of this system over the traditional mechanical voting machines is that it preserves the actual markings of the voter on the ballot. This is in contrast to the mechanical voting machines in which, if an error occurs, there is no original ballot which may be inspected to ascertain directly the intent of the [voter], . . . The same scanning devices were used to conduct the canvass on election night and the recanvass. Assuming the new scanning machines were reading ballots in the same manner on election night and during the recanvass, the result would have been the same.” Gejdenson alleged, however, that there had been a reduction in the number of votes cast for each candidate as recorded on election night and during the recanvass: for Gejdenson, a reduction of twenty-two votes; for Munster, a reduction of nine votes; and for Bingham, a reduction of six votes. Gejdenson asserted that neither during the election night canvass nor during the recanvass did the election officials themselves read the ballots, which were available “to determine the actual count of votes cast for each candidate.” Accordingly, Gejdenson requested that we direct that all of the ballots in Norwich be [617]*617recounted “by election officials and not by the new counting devices,” in order “to ascertain for which candidates votes contained thereon are cast.”
At our first hearing, on November 29,1994, Gejden-son moved for an immediate manual recount of the Norwich ballots, to be conducted while we heard the evidence on the other issues in the case. Munster objected to the timing of the Norwich recount. He argued that to order a recount to begin immediately, before a decision about a general recount of the entire district, would be improper because “if you order a recount as to any town you must order a recount as to all towns because the purpose of the recount is to apply a uniform decision making methodology, uniform rules of counting, to all towns.” Munster contended that any rules established for the Norwich recount must “be applicable in any other recount situation in any other town [for which] this panel may ultimately order recounts to be held so that we have uniform rules applying for the recounts.”
After full argument on this issue, we ordered an immediate manual recount of all of the Norwich ballots, including the absentee ballots. We rejected Munster’s objections to an immediate recount for the following reasons.16
First, both parties had asserted the inaccuracy of the count of the Norwich ballots and had requested a manual recount. Therefore, they agreed on the underlying need for such a recount in order to guarantee the reliability of the results of the election. Second, Munster’s argument for uniformity of rules for all recounts was unpersuasive, because the Norwich ballots—those cast at the polls and those cast absentee—were unlike any other methods of voting or ballots in any other town [618]*618in the district. We were unable on November 29,1994, and are unable now, after we have heard all of the evidence and reviewed all of the challenged Norwich ballots, to perceive how particular rules for counting the Norwich ballots could apply to any other ballots or methods of voting in the district other than the general rule that the voter’s intent should be dispositive. Third, to have awaited the ultimate outcome of these proceedings before ordering the Norwich recount that both parties agreed was necessary would have engendered an extraordinarily difficult legal question, the answer to which no one could confidently supply.17
Pursuant to our order, the Norwich recount proceeded immediately in accordance with the following procedures. We appointed Professor Colin C. Tait, of the University of Connecticut School of Law, to supervise the recount.18 His responsibility was to ensure the fairness and accuracy of the recount. He was not authorized, however, to render any judgment as to whether a ballot would be challenged or called as a vote, or as to the merits of any challenge, the determination of which remained with us.
There were six recount teams, each consisting of a ballot caller and two talliers (election officials). The ballot caller, who was a moderator who had worked both the election canvass and the recanvass, and was, there[619]*619fore, familiar with the ballots, was chosen by both the Republican and the Democratic registrars of voters in Norwich. The talliers were persons who had been absentee ballot counters at the election canvass and recanvass, or other election officials who had worked the election canvass and recanvass, and were, therefore, familiar with the ballots. The parties were allowed observers for each team. Only election officials handled ballots, which were counted by voting precincts.
Before a ballot was called as voted for a candidate or, in some instances, as a “no vote,” the caller afforded the observers a reasonable opportunity to inspect the ballot and to state whether a challenge would be made. Challenged ballots were not called as voted for any candidate. They were segregated and appropriately noted by the caller, who delivered them to Professor Tait. Unchallenged ballots were called as votes cast for one of the candidates or as “no votes,” as the case may have been. The talliers kept a record of the votes, of the “no votes,” and of the challenges, frequently comparing their tallies in order to avoid mistakes.
The recount took place in Norwich on December 1 and 2,1994. The results of the recount were as follows. Gejdenson received 5687 unchallenged votes, Munster received 3793 unchallenged votes, and there were seventy-three challenged ballots.19 In court, we examined each of the seventy-three ballots, and the parties stated their positions with respect to each ballot. In addition, the parties have had a full opportunity to present further argument on these challenged ballots, both in their posttrial briefs and in final oral argument. We turn, therefore, to our disposition of these seventy-three challenged ballots.
[620]*620The parties differ fundamentally concerning the proper legal standard that we should employ in determining the disposition of these ballots. Munster argues that, in order for a ballot to be voted for a particular candidate, the voter must have drawn a line that touches both ends of the interrupted or blank portion of the arrow. As Munster phrases it in his posttrial brief, we ought to “require a line touching both ends of the arrow in order to vote . . . .’’He does not argue that the width or intensity of the voter’s connecting mark be the same as the preprinted portions of the arrow; any width or intensity will suffice. In other words, in his view the voter must have completed the arrow in the sense that a discernible pen or pencil line connects both ends of the interrupted portion of the arrow. If, therefore, there is such a line that connects both ends of the arrow, the vote should be counted for that candidate. If, however, the line does not connect both ends but, for example, connects with one of the interrupted ends of the arrow but ends short in some discernible measure of the other end of the interrupted portion of the arrow, the vote is to be considered invalid and to be recorded as a “no vote.” In Munster’s view, therefore, our function is not to attempt to discern, to the extent reasonably possible, the voter’s intent in making the marks that he or she made, based upon all of the available evidence disclosed by the ballot. Our function, instead, is to examine the ballot in order to determine whether the voter complied strictly with the instruction on the ballot: “To vote, complete the arrow [depiction of completed arrow] at the RIGHT of the candidate’s name of your choice.”
Gejdenson argues, to the contrary, that our function is to determine, to the extent reasonably possible, the intent of the voter in making the marks that he or she made on the ballot, in light of all of the available evidence disclosed by the ballot. We agree with Gejden-son for three reasons.
[621]*621First, the process of voting, whether by mechanical machine of the kind traditionally used in this state, by our traditional absentee ballot, or by paper ballot to be electronically read, is essentially the process by which a voter expresses his or her intent that a particular candidate represent the voter in the office in question, subject, of course, to the legal principles governing the voting process. That expression of intent is accomplished through the means supplied by the state for that purpose, whether those means are a mechanical machine of the kind traditionally used in this state, our traditional paper absentee ballots, or the marksense demonstration process used in Norwich. Similarly, the process of counting votes, irrespective of the means supplied to the voter for the purpose of voting, is the process of tabulating the individual and collective expressions of the voters’ intentions, as disclosed by the particular means supplied for that purpose, and subject, of course, to the legal principles governing the voting process. Thus, in our view, voting and counting votes mean, respectively, expressing intent and tabulating those expressions of intent in accordance with the legal principles governing those processes. Whatever the process used to vote and to count votes, differences in technology should not furnish a basis for disregarding the bedrock principle that the purpose of the voting process is to ascertain the intent of the voters.
Second, the instructional materials issued by the secretary of the state to the moderators in Norwich support the conclusion that the ultimate goal of the demonstration project is to determine the individual and collective expressions of the voters’ intentions. In the “Moderator’s Handbook—1994 Marksense Demonstration Project,” the secretary of the state instructed the moderators as to how to treat a “blank or unreadable ballot,” namely, “one on which the machine can not read a single voter for a single office. It might occur [622]*622if someone intentionally submitted a ballot, but did not vote for anyone; however, it is much more likely that the voter has incorrectly recorded every one of his votes on the ballot (e.g., circled the names, used X’s or checks that the machine could not pick up, or used the wrong kind of pen or pencil).” In such a case, the machine would reject the ballot and the moderator was instructed to review the marking instructions with the voter and encourage him to obtain a new ballot and remark it. If that effort was unsuccessful, however, the voter “should be told that the ballot will be hand counted after the polls close, but if the election officials are unable to determine what he meant by his markings, some or all of his votes could be lost.” (Emphasis added.) The voter was then to be instructed to deposit his ballot in the “auxiliary bin,” in which such ballots were stored for hand counting after the polls closed.
Similarly, the secretary of the state’s “Recanvass Manual—1994 Marksense Demonstration Project” instructs that, before the recanvass officials run the previously machine counted ballots through the machine at the recanvass, “the ballots should be scanned for any defects or marking errors which could lead the machine to misread the ballot. (See Marksense Absentee Ballot Manual for a discussion of such errors and defects.) If any such errors or defects are found, the ballot should be set aside for hand counting of the races involved in the recanvass.”
The “Absentee Ballot Manual—1994 Marksense Demonstration Project,” referred to above, consistently emphasizes the importance of ascertaining the voter’s intent. Under the demonstration project, absentee ballots were run through the machine for counting in the same manner as the ballots of those voters voting in person. The absentee ballot manual contains instructions regarding blank ballots. These were bal[623]*623lots that were marked in such a way that the machine could not read any votes, because the voter made such “marks as circles, checks or ‘XV near the candidate’s name instead of filling in the vote indicator.” The manual instructed the election officials as follows: “Such ballots should be set aside for hand counting.” (Bold print in original.) Immediately below this instruction, the manual states: “Some ballots will have to be hand counted. The rule for counting ballots is that the intent of the voter governs. If the ballot is properly marked, the voter’s intent is clear. Many ballots are not properly marked. The statutes provide rules for determining the intent of the voter when the voter has incorrectly cast his ballot.” (Bold print and emphasis in original.)
It is clear, therefore, that, if an absentee voter failed to comply with the voting instructions, the process of hand counting the absentee ballots required a search for the intent of the voter. It is equally clear that the requirement of such a search is rooted in our statutes governing the counting of absentee ballots. See General Statutes § 9-150a (j).20 Because the Norwich absentee ballots were essentially the same as the Norwich ballots used by voters who voted at the polls, and [624]*624because the absentee ballots were designed to be run through the same counting technology, the conclusion is inescapable that the demonstration project contemplated the same search for the intent of the voter when the election officials were hand counting ballots of voters who had voted at the polls. The conclusion is equally inescapable, therefore, that the manual count of all of the Norwich ballots that the parties requested and that we ordered should also be governed by a determination of the intent of the voter as disclosed by his or her ballot. Any other conclusion would have the biz-zare result of requiring us to discern the intent of absentee voters, while requiring us to ignore the intent of voters who voted at the polls, despite the fact that both sets of voters used essentially the same ballot and voting technology.
Third, we have long adhered to the principle that ballots should, where reasonably possible, be read so as to effectuate the expressed intent of the voter, so as not unreasonably to disfranchise him or her. “Where the legislature in express terms says that a ballot shall be void for some cause, the courts must undoubtedly hold it to be void; but no voter is to be disfranchised on a doubtful construction, and statutes tending to limit the exercise of the ballot should be liberally construed in his favor. Unless a ballot comes clearly within the prohibition of some statute it should be counted, if from it the wish or will of the voter can be ascertained. State v. Bossa, 69 Conn. 335, 341, 37 A. 977 [1897]; Flanagan v. Hynes, 75 Conn. 584, 588, 54 A. 737 [1903]; Moran v. Bens, 144 Conn. 27, 32, 127 A.2d 42 [1956].” (Emphasis added; internal quotation marks omitted.) Scully v. Westport, 145 Conn. 648, 651-52, 145 A.2d 742 (1958). We see no reason to conclude that the legislature in enacting No. 94-225 of the 1994 Public Acts, and the secretary of the state in implementing it, intended either to depart from this fundamental prin[625]*625ciple or to subvert the democratic process designed to ascertain and implement the will of the people.
Munster’s argument to the contrary would require us to ignore the nature of the voting and vote counting process, the clear implications of the demonstration project materials,21 and this longstanding democratic principle of election law jurisprudence. It would substitute for the common sense task of determining the voter’s intent a mechanical determination of whether the voter’s mark touched both ends of the interrupted portion of the preprinted arrow. It would convert the printed instructions on the ballot, which were obviously intended to make it easier for the marksense machine to read the ballot and to avoid difficulties in that reading process, into the exclusive manner by which all ballots could legitimately be cast. It would elevate the form of those instructions over the substance of the voting process itself, and would ignore the entire thrust of the instructions to the moderators regarding manual counting of any ballots rejected by the machine. Furthermore, it would subject to an impermissible level of scrutiny, and would risk disfranchising, the elderly, the infirm, the physically or visually disabled and those with marginal literacy skills, who are those most likely to have made the kinds of marks that Munster’s proposed test would disqualify.
Such a test, namely, whether the voter’s mark touched both ends of the interrupted portion of the arrow, is not what was required by a manual count on election night, is not what was required by a manual [626]*626count during the recanvass, and is not what should be required of the manual recount that we ordered. Such a test lacks support in law, reason, experience or sound public policy.
With the proper test in mind, therefore, namely, a search for the expressed intent of the voter, we have scrupulously examined all seventy-three of the challenged ballots. We find as follows with respect to those seventy-three ballots: nineteen ballots should be counted for Munster;22 fifty-two ballots should be counted for Gejdenson;23 and two ballots should, by [627]*627agreement of the parties, be counted as “no votes.”24
Excluding the seventy-three challenged ballots, Munster received 3793 votes and Gejdenson received 5687 votes in the Norwich recount that we ordered. Adding in the results of our disposition of the seventy-three challenged ballots, those totals are changed to 3812 votes for Munster, and 5739 votes for Gejdenson.
This does not end our task, however. In order to determine whether any party gained or lost votes as a result of our Norwich recount, we must compare these totals to the total votes for each party according to the number of votes shown on the return to the secretary of the state.
According to the secretary of the state’s return of the Norwich recanvass ordered by her, Munster had received 3802 votes and Gejdenson had received 5711 [628]*628votes. Comparing those figures to the results of the recount that we ordered, we find that, as a result of that recount, Munster’s total votes from Norwich should be increased by ten votes (3812 minus 3802), and Gejdenson’s total votes from Norwich should be increased by twenty-eight votes (5739 minus 5711). This means that, as a result of the Norwich recount, there was a net gain of eighteen votes (twenty-eight minus ten) in favor of Gejdenson.
To summarize, as a result of the Norwich recount that we ordered, and through our careful review of the seventy-three Norwich ballots challenged by the parties, we have determined that, before any of Munster’s other claims are considered, Gejdenson has a preliminary lead of twenty-two votes, which is comprised of his four vote lead in the secretary of the state’s recan-vass results plus the net gain of eighteen votes as a result of the Norwich recount. The secretary of the state’s recanvass results showed Gejdenson with 79,160 votes and Munster with 79,156 votes. Adjusting for the Norwich recount, we now show Gejdenson with a preliminary total of 79,188 votes (an increase of twenty-eight votes), and Munster with a preliminary total of 79,166 votes (an increase of ten votes).
B
MUNSTER’S OTHER CLAIMS
Munster’s original petition challenged the results of the election in several towns other than Norwich, and in several other respects. Because of the time constraints mandated by § 9-323, which made it difficult to undertake a full investigation of alleged irregularities before the drafting and filing of his original petition, Munster’s original claims were broader than his final claims. During the hearings, Munster represented to the panel that his final claims, as ultimately required to be responded to by Gejdenson and to be determined [629]*629by us, would be only those presented in his posttrial brief. We turn, therefore, to the claims contained in that brief.
Munster’s claims fall into four categories, which can be characterized as follows: (1) improper counting of certain identified ballots; (2) irregularities in absentee ballot handling; (3) disparate treatment of similar cases; and (4) other irregularities. The evidence regarding these claims, Munster asserts, requires that we order a new election.
We disagree. We conclude, to the contrary, that considering all of the evidence in the light of the applicable law, there is no merit to Munster’s claim for a new election. We conclude further, based on the evidence presented to us and the facts we find to have been proven by a preponderance of that evidence,25 that: (1) those few irregularities that we find to have been proven are wholly insufficient to undermine the election; (2) it would be inconsistent with the applicable law, and with the sound democratic principles that underlie that law, to order a new election; and (3) Gejden-son was duly elected on November 8, 1994, as the United States Representative for the Second Congressional District of Connecticut.
Improper Counting of Ballots
Munster claims that certain votes that were cast in the towns of Waterford, Killingly and Bozrah constitute voting irregularities. We find most of these claims to be unproven.
[630]*630i
Ballots discovered at recanvass
In Waterford, twelve absentee ballots were counted at the recount, seven for Gejdenson and five for Munster, even though they had not been counted on election day. Munster claims that these votes should not have been counted. We disagree.
The facts are that these ballots had been properly executed, properly delivered to the town clerk and subsequently properly delivered by her to the absentee ballot counters for the noon count of absentee ballots on election day. By some oversight, however, the envelopes containing these ballots were not opened by the counters on election day and, therefore, the ballots were not counted. Instead, after the count, the unopened envelopes containing these ballots were returned to the town clerk in the appropriate sealed depository envelope designated for the purpose of maintaining election materials. The town clerk retained these envelopes in the sealed depository envelope, along with all of the absentee ballots that had been counted and the corresponding absentee ballot envelopes, until the day of the recanvass.
At the recanvass, it was discovered that these envelopes had not been opened, and that the absentee ballots in them had not, therefore, been counted on election day. In accordance with the ruling of the election officials, the envelopes were opened and the ballots counted.
Munster argues that the votes cast by these ballots should be invalidated because the ballots were not originally counted on election day. He contends that, pursuant to § 9-311a; see footnote 4; and General Statutes [631]*631§ 9-311,26 “the universe of ballots subject to recanvass on recanvass day is limited to those ballots that were initially canvassed on Election Day, either counted or [632]*632rejected.” (Emphasis in original.) Thus, he argues, “[t]he issue upon recanvass is, therefore, nothing more than recounting those absentee ballots considered by [633]*633election officials on Election Day, with or without changes in rulings accepting or rejecting ballots that had been questioned.” (Emphasis in original.) We disagree.
Section 9-311a provides for the mandatory recanvass that was conducted in this case. Section 9-311 provides the general procedures for conducting a recanvass. There is nothing in either the language or the purpose of either statute to support Munster’s cramped reading of these statutes.
The language of the statutes simply refers to a “recanvass.” The statutes’ purpose is to ensure the true and most accurate count possible of the votes for the candidates in the election. We can conceive of no rationale upon which the legislature, in enacting those statutes, would have intended that the wholly valid ballots of absentee voters, who had complied with all legal requirements for absentee voting, and whose ballots had been properly received by the town clerk and properly delivered to the ballot counters for counting, should nonetheless be excluded from the recanvass solely because the counters overlooked them on election day. Indeed, such a mistake in counting on election day is precisely the kind of mistake that a recanvass is designed to remedy.
A different conclusion would, without any basis in law or reason, disfranchise those twelve voters. This would fly in the face of our longstanding principle that “no voter is to be disfranchised on a doubtful construc[634]*634tion, and statutes tending to limit the exercise of the ballot should be liberally construed in his favor.” (Internal quotation marks omitted.) Scully v. Westport, supra, 145 Conn. 651-52. We decline to adopt such a bizarre construction of our recanvass statutes.
In sum, therefore, with respect to the set of claimed irregularites characterized as ballots discovered at recanvass, we conclude that all twelve ballots were properly counted at the recanvass.
Ineligible voters
Munster claims that, in Killingly, three voters, Jennifer O’Leary, George Katsikis and Tasia Katsikis, were improperly permitted to vote on election day even though their names had been deleted from the voter lists. He also claims that, in Bozrah, one voter was improperly permitted to vote on election day even though the address given for him on the voter list was an uninhabitable dwelling. We agree with the first of these claims, but not with the second.
Munster’s legal claim about the voters in Killingly is that they were improperly restored to the voter lists in violation of the provisions of General Statutes § 9-42.27 Thus, he argues, their three votes were invalidly counted and, because it cannot be determined [635]*635for whom they voted without infringing on the voters’ rights to a secret ballot under article sixth, § 5, of the Connecticut constitution,28 they constitute significant irregularities in the voting process. We agree that these voters should not have been permitted to vote.
The facts regarding the claims about voter restoration are as follows. Jennifer O’Leary registered to vote in Killingly in August, 1992, listing her address as 31 Shepard Hill, where she resided with her parents. In May, 1994, pursuant to General Statutes § 9-3529 the [636]*636Killingly registrars of voters sent her a canvass card that she returned in July, indicating that she had moved to 93 School Street, Putnam. Accordingly, the registrars removed her name from the Killingly voting list.
[637]*637O’Leary returned the canvass card without thinking through the matter. She did not register to vote in Putnam. She had rented an apartment there because it was convenient to her place of work. Nonetheless, she still considers 31 Shepard Hill Road, Killingly, to be her permanent address and home. She maintains a bedroom and her wardrobe there, she keeps her pets there, her car is registered there, her driver’s license indicates that as her residence, and most of her mail goes there.
When O’Leary went to Killingly to vote, she persuaded the election workers that she continued to be a resident of Killingly even though her name was no longer on the voting list. The election workers wrote her name on the voting list at 31 Shepard Hill Road and permitted her to vote. The election workers acted, however, without the consent of either registrar of voters, and without requesting that O’Leary fill out a voter registration card. After the election, the registrars of voters restored her to the voting list at the Shepard Hill address.
The facts about George and Tasia Katsikis are that, although they had been registered to vote in Killingly, they had indicated that they had moved to 10 Bunny [638]*638Lane, Danielson,30 in reply to canvass cards mailed to them by Rita LaBelle, one of the Killingly registrars of voters. Their new address is on a street that, although it has a Danielson, and thus Killingly, post office, is actually across a bridge and is part of the town of Brooklyn. Accordingly, in July, 1994, the registrars of voters removed their names from the Killingly voting lists. On election day, the Katsikises went to the Killingly voting place for Adelaide Street, and satisfied the election workers, including LaBelle, that they lived at 95 Adelaide Street. Their names were written on the voting list, with 95 Adelaide Street as their address, and they were permitted to vote. They were, however, restored to the voting list without the consent of Emily Harrington, the other registrar of voters.
Munster’s claims regarding O’Leary and the Kat-sikises rest on the provision of § 9-42 that requires that “no name shall be added to the list on election day, under the authority conferred by this section, without the consent of both registrars.” See footnote 27. He argues that this provision is mandatory, and that the lack of that joint consent prohibited any election officials from adding O’Leary’s and the Katsikises’ names to the voting lists on election day and, thus, from permitting them to vote. We agree.
Section 9-42 provides the procedure pursuant to which a voter, whose name was stricken from a voter list, may have his or her name restored to the list. It contains a specific provision, however, phrased in what is ordinarily considered to be mandatory language, for such a restoration on election day: the consent of both registrars of voters. Both registrars were available during election day, albeit perhaps not right then and there. The inconvenience of awaiting their consent, however, is not enough to overcome the important policies behind the statutory provision.
[639]*639Because those registrars are each selected from the two major parties—Democratic and Republican—the manifest purposes of this statutory limitation are to protect against one party taking improper advantage of the other, and to ensure that the members of each party are confident that their interests are protected against such an event. We cannot consider this provision as anything other than mandatory. We therefore conclude that O’Leary and the Katsikises were improperly added to the voting lists on election day and that they were improperly permitted to vote.
Munster also claims that, in Bozrah, George Gager was permitted to vote in violation of § 9-35. Gager was on the Bozrah voting list with an address of 15 Gager Road. On election day, he satisfied the poll checkers in charge of that list that he was a resident of Bozrah, showed appropriate identification, and was permitted to vote. In fact, 15 Gager Road is, and has been for three years, an unoccupied, uninhabitable, burned out house.
Munster argues that, like the votes of O’Leary and the Katsikises in Killingly, Gager’s vote should be considered invalid because “the voter was not living where the list showed him as living, and there is no evidence he was living anywhere else in Bozrah.” We disagree. Unlike the cases of O’Leary and the Katsikises, there is no evidence here that anyone made a transfer of Gager’s registration to a new address on the registry list. There was, therefore, no proof of a violation of § 9-42 by such a transfer of registration without the consent of both registrars. Moreover, Munster, who had the burden of persuasion regarding any voting irregularity that he claimed, did not prove that Gager did not reside in Bozrah on election day or was not registered in Bozrah on election day. Although his specific address as disclosed by the list may have been inaccurate, we [640]*640conclude that any irregularity in permitting him to vote under these circumstances was de minimis.
In sum, therefore, with respect to this set of claimed irregularities characterized as ineligible voters, we conclude that Munster has established that three voters in Killingly were permitted to vote who should not have been, and that Munster has failed to prove that the challenged vote in Bozrah was improperly cast.
Irregularities In Absentee Ballot Handling
Under this heading, Munster raises two different categories of claims of irregularities regarding absentee ballots. The first is the claimed lack of proper signatures of certain town clerks on outer envelopes containing absentee ballots that had been properly delivered to those town clerks. The second consists of a series of other claimed irregularities in the documentation required by the statute for absentee ballots.
The signature of the town clerk (General Statutes § 9-140c [a])31
Chapter 145 of the General Statutes §§ 9-133f through 9-159r, governs the process of voting by absentee bal[641]*641lots. Insofar as these statutes are relevant to this case, those statutes require, in general, that the following procedures be followed.32 Moreover, the evidence in this case persuades us that, except in those few discrete instances of irregularity noted later in this opinion, the election officials throughout the district followed the statutory requirements with admirable scrupulousness.
An applicant for an absentee ballot must deliver his or her application to the town clerk on a form prescribed and furnished by the secretary of the state. General Statutes § 9-139a (a). Upon receipt of the application, the town clerk, after ascertaining that the applicant is a registered voter in the town, must deliver an “absentee voting set” to the applicant. The set consists of an absentee ballot, inner and outer envelopes for the ballot’s return, and instructions for its use. General Statutes § 9-140 (d). The outer envelope must contain a serial number, and the town clerk must maintain a log, identifying the name and address of the applicant and the serial number of the outer envelope delivered to the applicant. General Statutes § 9-140 (a).
Upon receipt of the set, the absentee ballot voter must sign the form printed on the inner envelope, which, in accordance with General Statutes § 9-137,33 [642]*642constitutes a statement under the penalties of false statement that the applicant is entitled to vote by absentee ballot. General Statutes § 9-140a.34 Pursuant to General Statutes § 9-140a, the absentee ballot voter then inserts the ballot into the inner envelope, which the voter in turn inserts into the outer envelope for return to the town clerk.
Upon receipt of the outer envelope, the town clerk logs it in, checking it against his or her log of absentee ballot sets sent out. The town clerk also is required to endorse on the outer envelope, over his or her signature, the date and time of its receipt. General Statutes § 9-140c (a). The town clerk is required to execute an affidavit attesting to the accuracy of all of his or her endorsements and, at the close of the polls on election day, to deliver that affidavit to the head moderator, who in turn endorses on the affidavit the time of receipt. The head moderator later returns the affidavit to the town clerk after the counting of the absentee ballots. General Statutes § 9-140c. The town clerk [643]*643may not open any outer envelope, but retains all such envelopes securely in his or her office until they are ready for counting on election day.
The evidence before us disclosed that the counting procedure, as required by General Statutes §§ 9-140c, 9-150a and 9-150b,35 proceeds along the following gen[644]*644eral lines. First, the town clerk delivers the outer envelopes to the registrars of voters, who check the voters’ [645]*645name against the official checklist used at the polls on [646]*646election day and indicate thereon that the voters have [647]*647voted by absentee ballot.36 After this checking procedure is completed, the town clerk seals the unopened package of envelopes and retains them in a safe place. General Statutes § 9-140c (d).
At the appropriate time for counting, the town clerk delivers the sealed package of unopened envelopes to the ballot counters. General Statutes § 9-150a (b). The ballot counters open the package, remove and open the outer envelopes, remove the inner envelopes from the outer envelopes, and count the outer and inner envelopes, checking the two numbers against each other. They also report the total number of absentee ballots to the moderator. If an inner envelope has not been signed by the voter as required by § 9-140a, constituting a statement under penalties of false statement that the voter is entitled to vote by absentee ballot, the inner envelope is not opened; it is replaced in its outer envelope, the outer envelope is marked “rejected,” and the reason for the rejection is endorsed thereon by the counters. General Statutes § 9-150a (d).
The inner envelopes are then removed from the remaining outer envelopes, and all of the envelopes, including those rejected, are placed in a special “depository envelope” prescribed by the secretary of the state. The counters carefully seal these depository envelopes with nonreusable tape, endorse them appropriately, and deliver them to the moderator, who ultimately [648]*648delivers them to the town clerk. The town clerk maintains these depository envelopes, unopened, in a secure place for possible use in a recanvass. General Statutes § 9-150a (f).
At this point, the ballots, which have been removed from their inner envelopes, remain folded, so that the counters cannot ascertain their votes and match them with the envelopes from which they had been removed. Thus, the secrecy of the ballot is maintained. The counters then shuffle the folded ballots, so that they are in random order, before unfolding and counting them.
The counting proceeds in public, usually with representatives of each candidate observing so as to identify and, if necessary, question the call of an ambiguous ballot. If there is a question of the intent of the voter, the questioned ballot is submitted to the moderator, who makes the determination of whether the ballot should be cast and for whom, and endorses his or her decision on the ballot. General Statutes § 9-150a (k). After the ballots are counted, and the votes recorded by the counters, the counters seal them in other special depository envelopes in a fashion similar to the sealing of the depository envelopes for the outer and inner envelopes described above. General Statutes § 9-150a (m). These depository envelopes are delivered to the moderator, who later delivers them to the town clerk for similar safekeeping and preservation. General Statutes § 9-150b (e), (f) and (g).
With this factual background in mind, we turn to Munster’s challenges to certain ballots on the ground that the town clerks in question did not sign the outer envelopes, as required by § 9-140c (a). This claim involves a total of 413 outer envelopes for absentee ballots: 403 in Stonington; 3 in Old Saybrook; 6 in Led-yard; and 1 in Norwich. We conclude that, of those 413 ballots, only three were improperly signed in, and the remaining 410 were properly processed.
[649]*649The facts concerning the 403 Stonington absentee ballots disclose a uniform signature pattern.37 Although Munster introduced only one of the outer envelopes into evidence, and although he recognizes in his posttrial brief that he had not claimed any additional irregularities based on an alleged lack of a proper signature on other envelopes, the town clerk testified, and we find, that all 403 outer envelopes were treated in the same fashion. We, therefore, consider Munster’s claim to apply to all 403 absentee ballots cast in Stonington.
Upon receipt of each absentee ballot outer envelope, the town clerk stamped it using her time stamp machine, with the following printed legend38 appearing on the outer envelope:
“RECEIVED FOR RECORD STONINGTON, CT.
94 NOV-8 AM 9:55
RUTH WALLER TOWN CLERK”39
In Old Saybrook, the evidence before jus discloses that three of the outer envelopes received by the town clerk were stamped as follows:
“RECEIVED OCTOBER 14 1994”
[650]*650In Ledyard, although the town clerk had a stamp facsimile of her cursive signature that she customarily affixed to the outer envelope upon receipt, the evidence before us discloses that six of the outer envelopes lacked that cursive facsimile, because they arrived during a period in which she was recovering from heart surgery and was out of the office. Her assistant inadvertently failed to affix the town clerk’s cursive facsimile. These envelopes were stamped as follows:
“RECEIVED FOR RECORD AT LEDYARD, CT.
94 OCT 26 AM 10:31
ATTEST: PATRICIA EARNS TOWN CLERK”
In Norwich, although the town clerk had a stamp facsimile of her cursive signature that she customarily affixed to the outer envelope upon receipt, the evidence before us discloses that one outer envelope lacked that cursive facsimile, and was stamped as follows:
“RECEIVED
94 NOV-7 AM 9:03
BEVERLY C. MULDOON TOWN-CITY CLERK NORWICH, CONN”
In deciding whether these various signature stamps comport with the requirements of our law, our point of departure is § 9-140c (a), which provides in relevant part: “The municipal clerk shall endorse over his signature, upon each outer envelope as he receives it, the date and precise time of its receipt.” Munster concedes that a stamp of the cursive signature of the town clerk is sufficient for compliance with the statute, on the theory that “[without implicating the laws prohibiting forgery, no one can go to a stamp store and purchase a [651]*651stamp of an incumbent town clerk’s signature; only the town clerk can do that. There is, therefore, a decided security and ballot integrity reason for the statute’s requirement that the unique and personal signature of the town clerks, actually inscribed or stamped, be used as the receipt for absentee ballot outer envelopes.” Munster argues, however, that none of the exemplars shown above complies with the statute.
We agree with Munster that the stamp on the three Old Saybrook outer envelopes does not comply with the statute. We disagree with his claim, however, with regard to the remaining 410 outer envelopes, namely, those in Stonington, Ledyard and Norwich. With respect to those, we conclude that the stamps used by the town clerks sufficiently complied with § 9-140c (a).
The provisions of § 9-140c (a) regarding the date and time of the town clerk’s receipt of an absentee ballot envelope, and the town clerk’s signature, are mandatory because they are designed to mitigate the risk of fraud that is inherent in the absentee voting process. Dombkowski v. Messier, 164 Conn. 204, 319 A.2d 373 (1972). That does not mean, however, that strict, as opposed to substantial, compliance with those provisions is required. Rather, there must be substantial compliance with the statutory requirements. Wrinn v. Dunleavy, 186 Conn. 125, 147, 440 A.2d 261 (1982); Dombkowski v. Messier, supra, 209.
What constitutes a “signature” has been the subject of legislative enactments and case law outside the election process arena. Compare, e.g., General Statutes § 7-380 (referring to a “facsimile of any signature”), with General Statutes Annotated § 42a-3-401, comment 2 (West 1990) (“[a] signature may be handwritten, typed, printed or made in any other manner”); see also 72 Am. Jur. 2d, Statute of Frauds § 358 (1974) (“[t]he general rule is that the signature may be affixed by a [652]*652stamp, or it may be typewritten or printed mechanically, if ... by signing in any of these methods the party whose signature is essential intends to authenticate the instrument as his act”). We believe that whether the printed name and title of the respective town clerks, in the exemplars displayed above, constituted substantial compliance with § 9-140c (a), must be determined by reference to the purpose of the statutory requirement, the role played by the requirement viewed in the context of the statutory scheme, the degree of adherence to strict compliance shown, and the basic policy against disfranchisement of voters who are not at fault for any lack of strict compliance.
The purpose of the signature requirement in § 9-140c (a) is to avoid fraud in the voting of absentee ballots. By requiring the town clerk to sign the outer envelope, the statute seeks to avoid the risk that an unauthorized person will somehow include an unauthorized absentee ballot among those validly sent and delivered.
The signature requirement, however, must also be viewed in the context of the entire statutory scheme. That scheme requires that outer envelopes be logged out and logged in, that the serial numbers on the envelopes be checked against those on the sets assigned to absentee voters, that the town clerk deliver to the head moderator an affidavit attesting to the accuracy of all the endorsements on the outer envelopes, and that the count of outer envelopes eventually be checked against the count of inner envelopes. Because this procedural rigor is a significant safeguard against fraud, the fact that the evidence before us reveals punctilious adherence to the remainder of these requirements informs our determination as to whether there has been substantial compliance with the statute. Furthermore, the [653]*653extent of deviation from strict compliance is also relevant in drawing an appropriate line between substantial and insubstantial compliance.
Finally, we must recognize the overarching policy that, in construing voting statutes, “no voter is to be disfranchised on a doubtful construction, and statutes tending to limit the exercise of the ballot should be liberally construed in his [or her] favor.” (Internal quotation marks omitted.) Scully v. Westport, supra, 145 Conn. 651-52. We, therefore, take into consideration whether the failure of strict compliance was due to the conduct of the voter or of someone not within his or her control.
Applying these considerations to these ballots, we conclude that all but the three Old Saybrook envelopes met the substantial compliance test. The stamp on the Old Saybrook envelopes is merely a generic date stamp and contains no indication, whether by hand signature, stamp facsimile or printed name and title, that it was received by the town clerk. Furthermore, there is no time of receipt indicated on the stamp, as required by the statute. General Statutes § 9-140c (a). The minimal adherence to the requirements of § 9-140c (a) evinced by the endorsements on the three Old Saybrook envelopes in question leads us to conclude that they do not substantially comply with the requirements of § 9-140c (a).
The remaining 410 ballots, however, did substantially comply with those requirements. Although a stamped facsimile of the town clerks’ cursive signature would arguably have been preferable, we cannot ascribe critical significance to the difference between such a cursive facsimile and the printed names and titles of the town clerks that were rendered on the envelopes by the town clerks’ time and date stamp machines. Neither of these types of stamps is readily available to the public.
[654]*654The fact that the envelopes display the town clerks’ printed name and title, rendered thereon by the town clerks’ machines, sufficiently avoids the risk that unauthorized ballots might be mingled with authorized ballots. Furthermore, all of these envelopes were handled with the procedural rigor that the statutory scheme, as a whole, contemplates. Moreover, such a printed name and title is sufficiently close in degree to strict compliance with the requirement of a “signature” that it draws a cognizable line between substantial and insubstantial compliance. Finally, the strong policy against disfranchising voters in our construction of voting statutes counsels against a holding that the town clerks’ substitution of their machine printed names and titles for a rubber stamp of their cursive facsimiles would effect such a disfranchisement.
In sum, we conclude that three of the 413 questioned absentee ballots cast in the towns of Stonington, Old Saybrook, Ledyard and Norwich should not have been counted on election night or in the recanvass. The remaining 410 ballots were validly cast and counted.
Other claimed absentee ballot irregularities
Under this category, Munster challenges the validity of five absentee ballots that were counted in Putnam, East Lyme, Stonington and Plainfield. We consider each claim individually, and conclude that only one of them is persuasive.
With respect to Putnam, Munster claims that a ballot should have been disallowed because it was returned to the town clerk in an outer envelope with a portion of the envelope cut out. The portion cut out was the printed form that the town clerk had filled in prior to delivering the ballot set to the applicant. That printed [655]*655form is required by General Statutes § 9-139,40 which requires that the “outer envelope shall also contain . . . blank spaces . . . upon which the municipal clerk, before issuance of the ballot and envelopes, shall insert the the applicant’s name, voting residence by street and number, voting district, [and] the date of the . . . election ... at which the ballot is to be cast
The evidence before us disclosed that, upon inquiry by the town clerk, the voter stated that she did not want the town clerk to know who she was. Nonetheless, the voter had also written on the outer envelope, in the place for a return address, her full name and address, and on the back of the envelope, above the cut out portion, her full address.
Munster claims this as an absentee voting irregularity because “§ 9-139 requires that when the outer envelope is returned to the Clerk, it contain the identification data the Clerk put on it when it was sent to the prospective absentee voter.” We disagree.
Section 9-139 constitutes a direction to the town clerk to insert the required information on the outer enve[656]*656lope. The purpose of this statutory requirement is to aid the town clerk’s record keeping by making it easier to match the incoming ballots with existing records. The statute does not require that, if the voter, as she did in this case, obliterates that particular part of the outer envelope, the voter be disfranchised, at least where the town clerk can identify the voter and match the voter with his records to determine that the ballot set was completed by the person to whom it was assigned. Because in this case there was sufficient information for the town clerk to make that determination, there was substantial compliance with § 9-139.
With respect to Colchester, Munster claims that the two absentee ballots of a husband and wife, Michael D. Pineault and Christine M. Pineault, should have been disallowed. Although each had applied for and had been sent individual absentee ballot sets, they both returned their inner envelopes in the outer envelope furnished to and identified for Michael D. Pineault. Therefore, Michael D. Pineault’s inner envelope was returned inside his own outer envelope, but Christine M. Pin-eault’s inner envelope was returned inside her husband’s outer envelope. Both ballots were counted. Christine M. Pineault testified, and we find, that she voted for Munster. Michael D. Pineault did not testify.
Munster argues that “§ 9-139 requires that the voter’s ballot be contained in a separate inner envelope enclosed in the outer envelope issued to that voter by the Clerk and containing the identification data the Clerk had placed on that outer envelope at the time of issuance.” Therefore, he asserts, both “ballots that came in this envelope should not have been counted.”
We agree with Munster that § 9-139 requires that each inner envelope be returned to the town clerk in its own outer envelope. That does not mean, however, that under these facts both absentee ballots were [657]*657improperly counted. Michael D. Pineault’s inner envelope was properly inside his outer envelope. Therefore, his ballot was properly counted.
Christine M. Pineault’s inner envelope, however, was not returned inside her outer envelope. Therefore, her ballot should not have been counted. This does not help Munster, however, because she voted for him. Because Gejdenson does not seek to have her disfranchised, we are not inclined to do so. In sum, we conclude that there was one voting irregularity in Colchester, but that under no circumstances could it have harmed Munster.
With respect to East Lyme, Munster claims that the ballot of an absentee voter should have been disallowed because the voter improperly identified himself on the ballot. The face of the ballot clearly indicated the voter’s intent to vote for Gejdenson, because he put an “X” in Gejdenson’s box. He also put an “X” in the box for the Republican candidates for governor and lieutenant governor, John G. Rowland and M. Jodi Rell. He then, however, crossed out that box, wrote “Delete” next to it, and signed his name underneath. He also put an “X” in the box for the Independence Party candidates for governor and lieutenant governor, Tom Scott and Glen R. O’Keefe. At the top of the ballot, he wrote: “Dear Sir, Please register my vote for Scott instead of Rowland. Thank you. All others remain the same.” He then signed his name.
Munster claims that this ballot should not have been counted for Gejdenson because it violated the proscription of § 9-150a (l)
[658]*658Section 9-150a (l) does not require the rejection of any ballot on which the voter has identified himself. Rather, it provides that “[n]o absentee ballot shall be rejected as a marked ballot unless, in the opinion of the moderator, it was marked for the purpose of providing a means of identifying the voter who cast it.” (Emphasis added.) The question, therefore, is one of the voter’s intent in marking his ballot. If his intent was. to provide “a means of identifying” himself, the ballot must be rejected. If, however, he had some other intent in marking his ballot, the ballot does not come within the proscription of § 9-150a (l).
Although the statute lodges the initial determination of the voter’s intent in the moderator; see also General Statutes § 9-150a (k); we are not persuaded by Gej-denson’s argument that, in our review of absentee ballots pursuant to § 9-323, we must give deference to the validity of the moderator’s ruling. We agree with Munster that we must determine de novo the voter’s intent in casting or marking an absentee ballot.
We have employed such a plenary standard in our appellate review of a trial judge’s finding regarding a voter’s intent in casting an absentee ballot. See Hurlbut v. Lemelin, 155 Conn. 68, 230 A.2d 36 (1967).42 Moreover, ordinarily the question of intent must be determined solely from the ballot itself, because that is all that the moderator or we have available. Because we are as skilled as a moderator in making such a determination and have before us all of the evidence that the moderators had before them, we see no practical or legal benefit to constraining our review and leav[659]*659ing in place a factual finding of intent if, in light of all the evidence available, we consider it to have been wrongly determined.
Applying this scope of review to the East Lyme ballot, we find that the ballot was properly counted because the message contained thereon does not violate § 9-150a (l). We find that the voter’s intent was not to furnish a means of identifying himself as the voter, but rather was to ensure that his change of marking on the gubernatorial line be understood as such, and that all of his other markings be understood to remain the same. Although it would have been preferable for him not to sign his name, taken in context his signature did not invalidate his ballot under § 9-150a (l).
With respect to Stonington, Munster claims that the absentee ballots of a husband and wife, Leo Besso and Esther Besso, should not have been counted because each returned his or her ballot to the town clerk in the other’s outer envelope. The evidence before us discloses that Leo Besso and Esther Besso applied for and were furnished separate absentee ballot sets. Leo Besso returned his inner envelope inside Esther Besso’s outer envelope, and affixed to the outer envelope a preprinted return address sticker with his name on it. He signed his name to the inner ballot in the appropriate place, and dated his signature. That outer envelope was returned to the town clerk in time for the counting of the absentee ballots, and was counted.43
Esther Besso returned her inner envelope inside Leo Besso’s outer envelope, and printed her name over his name on a similar preprinted return address sticker. [660]*660This outer envelope did not reach the town clerk, however, until November 14,1994, six days after election day. Therefore, Esther Besso’s ballot was not counted.
Although Munster raises a challenge to Leo Besso’s ballot that differs from the challenge he raised to the ballots of the Pineaults in Colchester, we cannot distinguish between the two cases. We agree with his claim that Leo Besso’s ballot should not have been counted because it did not arrive in his own outer envelope. See General Statutes § 9-139.
With respect to Plainfield, Munster claims that the absentee ballot of Clara Trahan should not have been counted because her marking of her ballot was not properly authenticated. The evidence before us discloses that Clara Trahan, who is a resident of and properly registered voter in Plainfield, but who lives at a convalescent home in Norwich, voted by absentee ballot pursuant to the statutory provisions for absentee ballot voting supervised by registrars of voting. See General Statutes §§ 9-159q and 9-159r.44 When she applied for her absentee ballot, she marked an “X” in the box [661]*661bearing the legend: “I am in an institution where registrars of voters supervise absentee voting.” Similarly, she signed the application with an “X.” The Plainfield town clerk forwarded an absentee ballot set to the Nor[662]*662wich town clerk under the appropriate supervised voting procedure. Trahan executed the ballot, dated the inner envelope, and signed the inner envelope with an “X.” Her “X” on the inner envelope is followed by the [663]*663signature of John H. White, a registrar of voters in Norwich, and his signature is followed by “ROV,” indicating his position as registrar of voters. Trahan’s ballot was counted.
Munster claims that Trahan’s ballot should not have been counted because it violates the provision of § 9-140a that “[a]ny absentee ballot applicant who is unable to write may cause his name to be signed on the form [on the inner envelope] by an authorized agent who shall, in the space provided for the signature, write the name of the applicant followed by the word ‘by’ and his own signature.” We disagree.
[664]*664First, we are not convinced that this statutory provision applies to this case. Trahan’s “X” constituted a valid signature by her. The law generally has long recognized that one who, for whatever reason, cannot sign his or her name in block or cursive letters, friay do so by affixing an appropriate mark. See generally 72 Am. Jur. 2d, Statute of Frauds § 359 (1974). White was not signing the inner envelope for her; he was simply signifying that, in his capacity as a registrar of voters, he had witnessed her mark as her signature.
Second, even if we apply the statute to this case, we conclude that there was no irregularity. White’s signature followed by his title, both of which were directly to the right of Trahan’s “X,” constituted substantial compliance with the statute.
In sum, therefore, with respect to the set of five other claimed irregularities in these identified absentee ballots, we conclude that there were two irregularites, only one of which could possibly have harmed Munster. The remaining three ballots were properly counted.
Different Treatment of Similar Cases
Under this category, Munster claims that, in four sets of instances, “disparate treatments] of similarly-situated matters” demonstrate significant electoral irregularities. Those four claims involve: (1) absentee ballots in Preston and Middletown that were allegedly treated differently from similar ballots in Old Saybrook, Ledyard, Stonington and Norwich; (2) an absentee ballot in Windham that was allegedly treated differently from a similar write-in machine vote in Tolland; (3) an absentee ballot in Brooklyn that was allegedly treated differently from similar absentee ballots in Old Lyme, Waterford, Plainfield, Old Saybrook, New London, [665]*665Windham and Middletown;45 and (4) an absentee ballot in Waterford that was allegedly treated differently from a similar absentee ballot in Windham. We agree with Munster’s second claim above, as indeed does Gej-denson, and conclude that Munster should be credited with one additional vote on the basis of the Tolland write-in vote. We disagree, however, with the other three claims.
Preston and Middletown v. Old Saybrook, Ledyard, Stonington and Norwich
These claims of disparate treatment all involve the treatment of absentee ballot envelopes in various towns with respect to the manner in which they were processed upon their receipt by the respective town clerks. Munster claims that there were two ballots rejected in Preston because they were not stamped as having been received by the Preston town clerk, although in fact they had been timely received by her. He claims that this is disparate treatment from one similar ballot counted in Middletown, three counted in Old Saybrook, six counted in Ledyard, and one counted in Stoning-ton. We agree with Munster that there were isolated inconsistencies. We disagree, however, that these inconsistencies have any legal significance.
The Preston town clerk testified, and we find, that she rejected two ballots because, although she had received them in her office in timely fashion, she had failed to stamp on the outer envelope the date and time [666]*666that she had received them. These envelopes were not introduced into evidence. We conclude that she acted properly in rejecting these ballots. See General Statutes § 9-140c (a) and our discussion of the Old Saybrook absentee ballots in part II B 2 i of this opinion.
The two Preston envelopes were not comparable, however, to the Middletown envelope. Whereas the Preston envelopes were rejected because there was no date or time of receipt noted on them, the Middletown envelope does have the date and time of receipt stamped on it. The Middletown envelope was rejected for a different reason, namely, the lack of the town clerk’s signature. Thus, with respect to Preston and Middletown, there was no disparate treatment of similar situations. Furthermore, contrary to Munster’s suggestion, the Preston and Middletown envelopes cannot be properly considered together, because they were rejected for different reasons.
The two Preston ballots, which were properly rejected by the town clerk for lack of a date and time stamp, were similar to the three Old Saybrook ballots, which were not rejected by the Old Saybrook town clerk and were counted. Thus, there was an inconsistency between these two towns in the treatment of a total of five ballots. We have determined, however, that the Old Saybrook ballots should not have been counted; see part II B 2 i of this opinion; and, therefore, any such inconsistency has been remedied by our declaration.
Comparing Preston and Middletown with Ledyard and Stonington, however, there was no disparate treatment of similar envelopes. The Preston and Middletown envelopes were properly rejected—the Preston envelope for lack of a date and time stamp, and the Mid-dletown envelope for lack of the town clerk’s signature. In contrast, we have concluded that the Ledyard and Stonington ballots were properly counted because the [667]*667town clerks’ stamp, which contained the dates and times of receipt and the names and titles of the town clerks, substantially complied with all of the statutory requirements.
li
Windham v. Tolland
In Windham, an absentee voter wrote in a vote for Gejdenson by writing “3C,” Gejdenson’s designation, on the write-in portion of his absentee ballot, rather than simply checking the “3C” box. This ballot was properly counted for Gejdenson, in accordance with the voter’s expressed intent. In Tolland, however, a voter using the voting machine, intending to vote for Munster by write-in rather than by lever, wrote in “3B,” Munster’s designation, on the write-in slot of the voting machine. This vote was improperly rejected. We agree, as does Gejdenson, that a vote indicated on the write-in slot of a voting machine should be gauged by the same standard as an absentee ballot: the expressed intent of the voter. Accordingly, one vote should be added to Munster’s total.
m
Brooklyn v. Old Lyme, Waterford, Plainfield, Old Saybrook, New London, Windham and Middletown
In Brooklyn, the moderator made a ruling regarding a questioned absentee ballot. He endorsed his ruling on the ballot as provided in § 9-150a. In the other towns, moderators made rulings on questioned absentee ballots but did not endorse their rulings thereon.
Although there are superficial inconsistencies in these instances, we conclude that these inconsistencies are de minimis and without legal significance. We regard the provision for endorsement by the moderator of his ruling on a questioned ballot as directory, rather than [668]*668mandatory. Its purpose is simply to make a record of those rulings so that the moderator, if later questioned in a recanvass or a subsequent judicial proceeding, may more easily recall the ruling and the reasons therefor. There was no evidence that, in the absence of these endorsements, the moderators’ rulings could not be properly challenged or reviewed. The provision, therefore, does not go to the essence of the validity of the ballot or the moderator’s task. Thus, although it is preferable for the moderators to comply with this directory provision, the fact that seven moderators did not do so has no legal significance.
iv
Waterford v. Windham
Munster claims that inconsistent determinations about what constitutes an overvote, which cannot be counted, occurred in Waterford and in Windham. He asserts that, in Waterford, “an absentee ballot with marks in the boxes for Gejdenson and Bingham was counted for Gejdenson . . . whereas in Windham an absentee ballot marked in the same boxes was correctly rejected as an overvote.” We find to the contrary.
This problem arises from the fact that, in Connecticut, when candidates have received the nomination of more than one party, the names of the candidate appear more than once on the ballot. For example, in this election, the name of Joseph Lieberman, a candidate for office of United States Senator, appeared on one line of the ballot as the nominee of A Connecticut Party (ACP) and on another line of the ballot as the nominee of the Democratic Party. This dual listing can lead some voters to mark the names of their preferred candidates at each place on the ballot where that name appears. Such ballots, crowded with multiple markings for particular candidates, may be difficult for some voters to verify. Voters may be confused because they have mis[669]*669takenly marked not just their preferred candidate’s name as it is listed by two different political parties, but may also have inadvertently marked the names of two different individuals who are candidates for the same office. Both of Munster’s allegations relate to such duplicate markings.
In light of the evidence before us and the complexity of the Connecticut ballots during this election, we conclude that the Waterford ballot was properly voted for Gejdenson. This ballot discloses that the voter voted for all of the Democratic candidates, including Gejden-son, with the exceptions of the Democratic candidates for governor and lieutenant governor, by filling in their boxes with black ink. The voter voted in similar fashion for the ACP candidates for governor and lieutenant governor, and for two charter questions that were on the ballot. The voter also registered his intent to vote for eight of the ten candidates on the ACP line, all eight of whom were the same candidates listed on the Democratic line for whom he voted by filling in their boxes. This intent was expressed, not by an “X” but by a clear check mark, in the eight boxes for those eight candidates on the ACP line.46
The two ACP candidates for whom the Waterford voter did not express an intent to vote were Bingham and the candidate for sheriff. Both of these candidates were different from the Democratic candidates for the same office. The voter left the sheriff’s box blank on the ACP line. In the Bingham box, there is a small vertical line that does not resemble in any way the check marks in the other eight ACP boxes. It is quite different in form, size and intensity. Applying our de novo scope of review to determine the voter’s intent, we find that he did not intend to register a vote for Bingham and, therefore, there was no overvote.
[670]*670In Windham, by contrast, the voter entered a definite “X” in each of the eleven ACP line boxes and in each of the twelve Democratic line boxes. These included a definite “X” in both Gejdenson’s and Bin-gham’s boxes. There is no way to differentiate between the “X” for Gejdenson and the “X” for Bingham, as indeed there is no way to differentiate between the various “X” marks entered in all of the other boxes on both the ACP and Democratic lines. Applying our de novo scope of review to this ballot, we find that this ballot was properly rejected as an overvote. See General Statutes § 9-150a (j) (1). Therefore, there was no disparate treatment between the Waterford and Wind-ham ballots.
In sum, with respect to this category of alleged disparate treatment of similar ballots, we conclude that there was one instance of such treatment and that, as a result, Munster’s total vote count must be increased by one. We also conclude that there was one other instance of disparate treatment that has in effect been remedied by our earlier declaration herein that three ballots in Old Saybrook should not have been counted. We conclude further that there were some isolated instances of inconsistent treatment that were without any legal significance. Finally, we conclude that there was no disparate treatment of similar ballots in the remaining instances claimed.
Other Irregularities
In this category, Munster asserts that other irregularities that occurred in two towns—East Lyme and Lebanon—substantiate his claim that the election results are fundamentally flawed. We find no such irregularities.
[671]*671i
Errors in tallying votes
Munster claims that in East Lyme “an absentee ballot counter who recorded 62 votes for Gejdenson erased it and put down 64 votes, based upon what his partner across the table had.” Our review of the evidence persuades us that this characterization is contrary to the facts.
In light of the credible testimony of Mary Grace Smith, who was the head moderator both on election night and during the recanvass, we find as follows. The recanvass, which is the relevant point in time for purposes of this claim, was conducted with scrupulous care and accuracy. The absentee ballot counters counted the ballots in packets of ten. Furthermore, periodically the counters compared their tallies with each other, and exhanged their packets for recounting, so that the ballots were counted not once, but twice. If at any time there was a discrepancy between the tallies, it was ultimately reconciled by the double count process. There is no credible evidence that an absentee ballot counter erased his count of sixty-two for Gejdenson and put down sixty-four based on the numbers that the other absentee ballot counter had recorded. Any erasure that may have occurred was simply a correction of a human error in recording, and consequently had absolutely no factual or legal significance with respect to the final count. Any change of recorded vote counts by any counter during the recanvass process could not have resulted, and did not result, in an inaccuracy in the final count. At the end of this scrupulous and painstaking process, both absentee ballot counters had exactly the same numbers for each candidate, and those numbers were correct.
[672]*672Ü
Discrepancies between the number of voters counted by hand and by machine
Munster claims that in Lebanon “the public counter on the machine in District One was three more than the number of people checked off as voting. The reasons given were in large part suppositious .... In District Two the total of the public counters for all machines was one more than the number checked off on the official checklist.” As a factual matter, we agree with the numerical discrepancies stated by Munster. We disagree with his characterization of the reasons given for the discrepancies as suppositious, and we conclude that the discrepancies were wholly without factual or legal significance.
The facts relevant to this claim are as follows. The “public counter” on a voting machine is a counter, visible on the side of the machine, that registers the number of persons who used the machine to vote. It is supposed to be “zeroed”—that is, the visible digit on the counter is to be brought to zero—by a voting machine mechanic before the voting begins. Before the polls open, the moderator checks that the public counter is set at zero, and at the end of the voting day the moderator reads and records on his tally sheet the ending number on the public counter.
Upon entering a voting machine, the voter pulls the red lever to the right in order to close the curtain on the machine. After voting by pulling down the levers above the candidates’ names, the voter pulls the red lever back to the left to open the curtain. If the machine operates properly, the counter advances by one digit each time a voter pulls the red lever back to the left. At the end of election day, the election officials com[673]*673pare the public counter number to the total number of voters who voted according to the official checklist.
Four voting machines were used in District One. During the day, one of the machines experienced a mechanical problem. The curtain failed to open when a voter pulled the red lever to the left, and the voter exited the machine through the closed curtain. The moderator determined that this voter had not moved the red lever sufficiently to the left to open the curtain, and asked one of the election workers who had not yet voted, but who was eligible to vote in that district, to use the machine to vote. The election worker did so, and upon moving the red lever to the right in order to close the curtain, the curtain opened instead. To correct the problem, the machine tender pressed the entry button and the voter pulled the red lever back to the left and again to the right in order to close the curtain, thereby registering a nonexistent additional voter.
In addition, later during the day a voter using the same machine reported a malfunction when he attempted to pull down the lever for governor. The moderator concluded that the voter had accidentally pulled down the write-in slot lever for governor, which automatically locks the pull down levers for that office. At that point, the machine was opened without registering the voter’s vote, and the moderator took the machine out of service and substituted a different machine, which the voter used.
At the end of election day, there was a three vote discrepancy between the total of the numbers on the public counters, 1117, and the total number of voters who, according to the official checklist, had voted by machine, 1114. The moderator concluded, and noted on his return, that two of the votes that form the discrepancy were caused by the machine adding a vote for each of the two times that it was opened when no [674]*674vote had been cast, as described above, and that the third vote “was probably due to the large crowd of voters at a separate time and someone must have gotten through without getting checked off the list.”
There is nothing suppositious, as Munster claims, about these inferences. We find to be rational and justifiable the inference that the machine malfunctions, as described above, account for two out of the three additional votes registered on the public counter over the number of voters indicated on the official check list. We also regard as rational and justifiable the inference that, during a busy and crowded time at the polling place, one voter went through the checkoff procedure without having his name checked off and subsequently voted. This one vote discrepancy was caused not by an improper vote being cast on the machine, but rather by an inadvertence, the absence of which would have resulted in the total from the official check list being 1115, rather than 1114.
In District Two, the total number of votes shown on the public counters of the four machines used was 1534. The total number of voters according to the official check list was 1533. There was, therefore, a discrepancy of one. The moderator concluded, and noted on his return, as follows: “The discrepancy of one voter is presumed to be an oversight by the official checkers during unusually heavy voting between the hours of five and 7:30 p.m.” As discussed previously, there is nothing suppositious about this inference.
In sum, under this category of alleged irregularities, we conclude that Munster’s claims were not proven. All of the explanations offered by the moderators for the numerical discrepancies were appropriately based on ordinary human experience or mechanical error. The processes of checking in voters at the polls and of voters using voting machines are human processes. People are [675]*675fallible, even when they strive to do their job carefully. Machines are also imperfect, and even when well maintained they may malfunction. Moreover, there is absolutely no proof, and no basis for a rational inference, that any person was improperly permitted to vote in either District One or District Two. We are entirely unpersuaded that these discrepancies, which are trivial in light of the total number of voters processed and, moreover, are wholly understandable, support Munster’s case in any way.
C
SUMMARY REGARDING IRREGULARITIES CLAIMED BY MUNSTER
Having fully evaluated all of the irregularities claimed by Munster, beyond those addressed by the Norwich recount, we reach the conclusion that, of all those claimed, there were only eight irregularities that had any legal significance: three voters were improperly permitted to vote; four absentee ballots were improperly counted; and one write-in vote for Munster was improperly rejected. With respect to this last instance, we conclude that Munster’s total vote count, including those votes added thereto as á result of the Norwich recount, must be further increased by one. After the Norwich recount, we showed a preliminary lead by Gejdenson of twenty-two votes. Adjusting that figure by the one vote increase for Munster, Gejdenson’s lead, after consideration of all of Munster’s claims, including the Norwich recount, is twenty-one votes.
With respect to the other seven irregularities, there was no proof regarding for whom those improper votes were cast. Even, therefore, if we were to indulge in an assumption that they were all cast for Gejdenson, the result of the election would not be changed, because under that hypothesis Gejdenson’s lead would be reduced to fourteen, and he would remain the winner. [676]*676See Scully v. Westport, supra, 145 Conn. 651; Armstrong v. Hartford, 138 Conn. 545, 551, 86 A.2d 489 (1952); Donovan v. Davis, 85 Conn. 394, 400, 82 A. 1025 (1912); State ex rel. Andrew v. Lewis, 51 Conn. 113, 124 (1883).47
In this connection, we address an issue that was raised in final oral argument. That issue is whether we should order a recount of the absentee ballots in the remaining fifty-three towns, other than Norwich, in the district. We conclude that there is no basis for such a remedy.
First, as a procedural matter, it is clear that, in his posttrial brief, Munster abandoned any claim for a recount of the absentee ballots in the other towns, and confined his requested relief to a new election, despite his claim in final oral argument that he was also requesting a recount of the absentee ballots in the remaining fifty-three towns.48 Nonetheless, rather than rest our rejection of such a possible recount on this [677]*677ground, we address on its merits the question of whether, on this record, there should be a recount of [678]*678the absentee ballots in the other fifty-three towns. We conclude that there should not be such a recount.
Section 9-323 provides in relevant part: “If sufficient reason is shown, [the panel] may order any voting machines to be unlocked or any ballot boxes to be opened and a recount of the votes cast, including absentee ballots, to be made. Such judges shall thereupon, in the case they, or any two of them, find any error in the rulings of the election official, any mistake in the count of such votes or any violation of said sections [9-355, 9-357 through 9-361, 9-364, 9-364a or 9-365, which are not involved in this case], certify the result of their finding or decision, or the finding or decision of a majority of them, to the secretary of the state before the first Monday after the second Wednesday in December.” The predicate for any action of the panel, therefore, including the ordering of a recount, is that “sufficient reason is shown.” Furthermore, such a “sufficient reason” must, insofar as this case is concerned, be based upon an “error in the rulings of [an] election official, [or a] mistake in the count of [the] votes.” There is nothing in this record to suggest that, in the other fifty-three towns of the district, there was any ruling of an election official or any conduct in the count of the votes that is not already adjudicated in this opinion. Thus, aside from any claims so adjudicated, there is nothing in the record to suggest any such improper ruling or mistake.
The legislature has wisely responded to the risks inherent in a close election by mandating an automatic recount, which was conducted throughout the district. Thus, there have already been two counts of the votes in those fifty-three towns—one on election night, and one during the mandatory recanvass ordered by the secretary of the state. At that recanvass, the election officials, accompanied by observers representing both candidates, were able to focus their attention on this [679]*679contest, without the haste required of the election night count. The recanvass was performed under necessarily less hurried and difficult conditions than the election night canvass, because there was more time available and the election officials were concerned, not with the results of all of the offices on the ballot, but only with this contest and that for the office of the secretary of the state. See footnote 4. The results of that recanvass have been duly returned to the secretary of the state. Munster had the opportunity to have observers at each of the fifty-three town recanvasses, and the opportunity of bringing to this panel any evidence of any erroneous rulings of election officials, of any mistakes in those counts, of any ballots that should not have been counted but that were counted, of any ballots that were counted but that should not have been counted, and of any other misconduct that could conceivably have occurred during the canvass and recan-vass in those towns. Other than the evidence he brought forward regarding specific instances of claimed irregularities in some of those towns, which we adjudicate in this opinion, he has produced no other evidence or claims of irregularities occurring in those towns.
There is, moreover, nothing in the recount of the Norwich ballots that we ordered that relates in any way to the absentee ballots in those fifty-three towns. Undoubtedly, the ballots and the counting technology used in the election and reviewed in our recount in Norwich were unlike anything used in the other fifty-three towns.
Moreover, Munster’s claim for another recount in the other fifty-three towns, in the absence of specific evidence regarding those towns beyond that already adjudicated herein, is contrary to existing precedent. In State ex rel. Andrew v. Lewis, supra, 51 Conn. 113, the petitioner, an unsuccessful candidate for mayor of New Haven, argued that the trial court should have ordered [680]*680a recount in those wards other than those covered by his petition and his proof. This court stated: “Under this statute a petitioner must allege the facts on which his claim to have been elected is based, and he must prove those allegations by [a] preponderance of [the] evidence as in all other cases. If he claims that mistakes had been made in the counting of votes in some particular ward or wards of the city, the correction of which would change the result, he must make out a prima facie case that such mistakes had been made before a recount of the votes in such ward or wards can be made. In such cases it will be presumed that the counting of votes in all the other wards of the city was correct, until the contrary appears.” Id., 123. This common sense principle was reaffirmed in Donovan v. Davis, supra, 85 Conn. 400, and has never been questioned.
Munster’s argument reduces to the assertion that, because this was a close election, and because there were some irregularities in some absentee ballots in the several towns with regard to which he did produce evidence, we should therefore infer that substantial irregularities occurred with regard to unchallenged ballots cast in those and other towns.49 There is no basis [681]*681for such an inference. On the contrary, we conclude that, after the election night canvass, and after the recanvass at which the counters were required only to count the ballots with respect to this contest and that for the office of the secretary of the state, the absence of any evidence in this case whatsoever regarding those fifty-three towns, other than the evidence concerning the claims already adjudicated, indicates that the recan-vass counts in those towns were accurate.
Thus, we reject the suggestion of Munster that a recount of the absentee ballots in the other fifty-three towns is necessary to dispel a cloud of doubt and uncertainty over this election. We have heard and have carefully considered all of the evidence presented to us regarding any claims of irregularity in voting and vote counting in the election. We have carefully considered, and adjudicate in this opinion, all the factual and legal claims made by Munster pursuant to which he challenges the result of the election. None of the evidence presented supports the inference that the claimed irregularities were just the tip of the iceberg. To the contrary, the fact that Munster has raised specific claims regarding the casting of ballots in many of the towns in the district leads us to conclude that he had ample opportunity to pursue all possible irregularities based on the canvass and recanvass already conducted.
The result of that adjudication is that, with respect to those fifty-three towns, there werefour absentee ballot irregularities, out of a total of approximately 8500 absentee ballots cast in those towns. There were an additional four nonabsentee ballot irregularities out of [682]*682a total of approximately 188,000 nonabsentee votes cast in those towns. The “cloud” that Munster perceives is no more than a chimera, conjured up out of rhetoric. It has no basis in fact, law or reason. It simply does not exist.
Indeed, these startling figures suggest more general factual findings and conclusions that we now address. Until this point in this opinion, we have been focusing our lens narrowly on the specific factual and legal claims of the parties, and on the specific evidence supporting and contradicting those claims. We cannot conclude this opinion, however, without also describing, in a more general way, the overall portrait that the evidence before us painted of the entire electoral process, and of the election officials who participated therein, in the district for this election.
The evidence persuades us that this was an eminently fair and accurate election. Beginning with the process of applications for absentee ballots, which began weeks before the election, through the many steps of the electoral process culminating in the recanvass and the Norwich recount, the election officials of the district worked with enormous diligence, honesty and success to ensure that every voter entitled to vote was afforded his or her opportunity, that every vote was counted truly, honestly and accurately, and that the candidate with the greater number of votes was declared the victor. This was consistently done without regard to party affiliation—whether of the voter or the official. At every step of the way, the election officials of the district worked impartially and with scrupulous attention to their official duties in order to ensure as much as possible that this election effected the will of the people of the district. This election was a model of our democracy in action.
[683]*683III
Conclusion
According to the returns of the mandatory recanvass, Gejdenson received a total of 79,160 votes. From that total, we subtract 5711, which was the total of his votes from Norwich according to that recanvass, for a preliminary subtotal of 73,449. To that figure we add 5687, which is the total of Gejdenson’s unchallenged votes from the Norwich recount, for a further preliminary subtotal of 79,136. To that figure, we add 52, which is the number of votes he gained from the 73 challenged ballots in the Norwich recount, resulting in a total vote count for Gejdenson of 79,188.
According to the returns of the mandatory recanvass, Munster received a total of 79,156 votes. From that total, we subtract 3802, which was the total of his votes from Norwich according to that recanvass, for a preliminary subtotal of 75,354. To that figure we add 3793, which is the total of Munster’s unchallenged votes from the Norwich recount, for a further preliminary subtotal of 79,147. To that figure, we add 19, which is the number of votes he gained from the 73 challenged ballots in the Norwich recount, for a further preliminary subtotal of 79,166. To that figure, we add 1, which is the write-in vote from Tolland for which he should have been credited, for a total vote count for Munster of 79,167.
Accordingly, pursuant to General Statutes § 9-323, and in accordance with the findings and conclusions rendered previously, we certify to the secretary of the state as follows: the number of votes cast for Sam Gej-denson for the office of the United States Representative for the Second District of Connecticut in the election of November 8, 1994, was 79,188; the number of votes cast for Edward W. Munster for the office [684]*684of the United States Representative for the Second District of Connecticut at the election of November 8, 1994, was 79,167. The winner of the election was Sam Gejdenson.
We vacate our earlier order directed to the secretary of the state, the treasurer and the comptroller of this state concerning their duties under General Statutes § 9-315.
We further order all of the town clerks of the fifty-four towns in the Second Congressional District to preserve all election materials used in that election until such time as the new United States House of Representatives that convenes in January, 1995, has, pursuant to article one, § 5, of the United States constitution, approved the qualifications of and accepted as a member of the House of Representatives a person as the United States Representative for the Second Congressional District of Connecticut.
In this opinion Peters, C. J., concurred.
Related
Cite This Page — Counsel Stack
653 A.2d 79, 231 Conn. 602, 1994 Conn. LEXIS 434, Counsel Stack Legal Research, https://law.counselstack.com/opinion/in-re-election-of-the-united-states-representative-for-the-second-conn-1994.