PAPER EVALUATION OF TRUNCATED ARROW-PER-LANE GUIDE SIGNS. WORD COUNT (5840): Abstract (248) + Text (4092) + Tables (5) + Figures (1) Authors:

Similar documents
Standard Highway Sign Border Specifications

H. T. Zwahlen, A. Russ, and T. Schnell 1

SECTION 7 RAMP TERMINAL SIGNS

Cipra D. Revised Submittal 1

Updating the Urban Boundary and Functional Classification of New Jersey Roadways using 2010 Census data

Land Navigation Table of Contents

TABLE OF CONTENTS. Page BACKGROUND...1 METHOD...3 PROCEDURE...6 RESULTS...8 READING TIME EVALUATION...13 DISCUSSION...17 REFERENCES...

AN INVESTIGATION INTO MEASURING THE CONTINUOUS WETTING RETROREFLECTIVITY VALUES OF PAVEMENT MARKINGS. A Thesis Proposal ADAM MATTHEW PIKE

A Driving Simulation Based Study on the Effects of. Road Marking Luminance Contrast on Driving Safety. Yong Cao, Jyh-Hone Wang

Safety and Performance Benefits Associated with the Use of a Spotter Mirror: Impact on Driver Lane- Change Planning and Execution

DATA DISAGGREGATION BY GEOGRAPHIC

Evaluation of fog-detection and advisory-speed system

Transportation and Road Weather

Transport Data Analysis and Modeling Methodologies

When is sign lighting needed? How much is needed?

C A R I B B E A N E X A M I N A T I O N S C O U N C I L REPORT ON CANDIDATES WORK IN THE CARIBBEAN SECONDARY EDUCATION CERTIFICATE EXAMINATION

Cost-Benefit Analysis of the Pooled- Fund Maintenance Decision Support System: Case Study

Geography and Earth Science. Unit 5

042 ADDITIONAL MATHEMATICS (For School Candidates)

Solar Powered Illuminated RPMs

$QDO\]LQJ$UWHULDO6WUHHWVLQ1HDU&DSDFLW\ RU2YHUIORZ&RQGLWLRQV

Technical Memorandum #2 Future Conditions

Confirmatory and Exploratory Data Analyses Using PROC GENMOD: Factors Associated with Red Light Running Crashes

III. FORECASTED GROWTH

EXAMINATION OF THE SAFETY IMPACTS OF VARYING FOG DENSITIES: A CASE STUDY OF I-77 IN VIRGINIA

Trip Generation Study: A 7-Eleven Gas Station with a Convenience Store Land Use Code: 945

Snow Removal Policy WARREN COUNTY ENGINEER S OFFICE (WCEO) WARREN COUNTY HIGHWAY DEPARTMENT. October 16, 2014

Expanding the GSATS Model Area into

Evaluation of Sign Retroreflectivity Measurements from the VISUAL Inspection Signs and Panels (VISUALISE) System

The students will have little or no prior knowledge of tires or their use. The teacher is familiar with the information being taught.

Effect of Environmental Factors on Free-Flow Speed

PHOENIZ, ARIZONA USA, SEPTEMBER 10, 2018 In keeping with a long tradition

Information for File # MMJ; Trunk Highway (TH) 7 / Louisiana Ave. Interchange Project

EXECUTIVE SUMMARY INTRODUCTION

Tsunami Sign Placement Guidelines

Deep Algebra Projects: Algebra 1 / Algebra 2 Go with the Flow

2010 Physics GA 3: Examination 2

Uncertainty in the Yellow Change Interval

محاضرة رقم 4. UTransportation Planning. U1. Trip Distribution

Contingency Tables. Safety equipment in use Fatal Non-fatal Total. None 1, , ,128 Seat belt , ,878

Texas Transportation Institute The Texas A&M University System College Station, Texas

2002 HSC Notes from the Marking Centre Geography

Puerto Rico Transportation Asset Management Plan

ILLINOIS ARMY NATIONAL GUARD Marseilles Training Center 1700 Army Road Marseilles, IL

Appendix B. Traffic Analysis Report

Objective: Construct a paper clock by partitioning a circle into halves and quarters, and tell time to the half hour or quarter hour.

DRAFT MANUSCRIPT. Title: Legibility of Text on Traffic Signs as a Function of Luminance and Size

A Cellular Automaton Model for Heterogeneous and Incosistent Driver Behavior in Urban Traffic

The webinar will begin shortly

1600 Prairie Drive , FAX

Objective: Recognize halves within a circular clock face and tell time to the half hour. (60 minutes) (2 minutes) (5 minutes)

CONGESTION REPORT 1 st Quarter 2016

AN INVESTIGATION INTO THE IMPACT OF RAINFALL ON FREEWAY TRAFFIC FLOW

Nominal Data. Parametric Statistics. Nonparametric Statistics. Parametric vs Nonparametric Tests. Greg C Elvers

6664/01 Edexcel GCE Core Mathematics C2 Bronze Level B2

Real-time, Adaptive Prediction of Incident Delay for Advanced Traffic Management Systems

The Next Generation of Traffic Management Systems

Deploying the Winter Maintenance Support System (MDSS) in Iowa

International Journal of Statistics: Advances in Theory and Applications

Encapsulating Urban Traffic Rhythms into Road Networks

Lecture 19: Common property resources

Rural Pennsylvania: Where Is It Anyway? A Compendium of the Definitions of Rural and Rationale for Their Use

Use of Spatial Interpolation to Estimate Interurban Traffic Flows from Traffic Counts

AS STANDARD

ADDRESSING TITLE VI AND ENVIRONMENTAL JUSTICE IN LONG-RANGE TRANSPORTATION PLANS

PHYSICS 111 SPRING EXAM 1: February 6, 2017; 8:15pm - 9:45pm

3.0 ANALYSIS OF FUTURE TRANSPORTATION NEEDS

DRAFT. New York State Testing Program Grade 8 Common Core Mathematics Test. Released Questions with Annotations

A Thesis Proposal. Agrawal, Ravi. Submitted to the Office of Graduate Studies of Texas A&M University

FEDERAL RESOLUTION PROPOSAL INTRODUCTION FORM

Trip Distribution Modeling Milos N. Mladenovic Assistant Professor Department of Built Environment

Research Article Headway Distributions Based on Empirical Erlang and Pearson Type III Time Methods Compared

The Sunland Park flyover ramp is set to close the week of March 19 until early summer

Figure 8.2a Variation of suburban character, transit access and pedestrian accessibility by TAZ label in the study area

CHAPTER 3. CAPACITY OF SIGNALIZED INTERSECTIONS

2010 HSC NOTES FROM THE MARKING CENTRE MATHEMATICS EXTENSION 1

Neighborhood Locations and Amenities

2016 VCE Specialist Mathematics 2 examination report

North Carolina s First Quadrant Left: History and Lessons Learned

Logistic Regression: Regression with a Binary Dependent Variable

Conservation of Mechanical Energy Activity Purpose

PHYSICS. Theory Examination. Administration and Marking Guide JUNE Secondary 5. Note

Best Practices and Techniques for Clearing Intersection Layouts

Statistical and Econometric Methods

BROOKINGS May

Rock Scaling Recommendations Logan Creek Drive Cut Slope

NATHAN HALE HIGH SCHOOL PARKING AND TRAFFIC ANALYSIS. Table of Contents

FINAL REPORT. City of Toronto. Contract Project No: B

Dunn County Snow Removal Policy

Time: 1 hour 30 minutes

Dust Storms in Arizona: The Challenge to Ensure Motorist Safety Jennifer Toth, P.E. Deputy Director for Transportation

AP CALCULUS AB 2011 SCORING GUIDELINES

TRB Traffic Flow on Freeway Upgrades. By Werner Brilon and Andrea Bressler

EVALUATION OF SAFETY PERFORMANCES ON FREEWAY DIVERGE AREA AND FREEWAY EXIT RAMPS. Transportation Seminar February 16 th, 2009

Information for File MVP RMM

Math 1020 ANSWER KEY TEST 3 VERSION B Fall 2018

Quality Pavement Markings

New York State Testing Program Grade 8 Common Core Mathematics Test. Released Questions with Annotations

FREEWAY WEAVING. Highway Capacity Manual 2000 CHAPTER 24 CONTENTS EXHIBITS

A-LEVEL STATISTICS. SS04 Report on the Examination June Version: 1.0

Transcription:

PAPER -0 EVALUATION OF TRUNCATED ARROW-PER-LANE GUIDE SIGNS Authors: Bryan J. Katz, Ph.D., P.E., PTOE Science Applications International Corporation 00 Georgetown Pike, F- McLean, VA 0 Phone: (0) - E-mail: bryan.j.katz@saic.com (Corresponding Author) Erin E. Dagnall, M.A. Science Applications International Corporation 00 Georgetown Pike, F- McLean, VA 0 Phone: () - E-mail: erin.e.dagnall@saic.com Mary Anne Bertola, M.A. Science Applications International Corporation 00 Georgetown Pike, F- McLean, VA 0 Phone: (0) - Fax: (0) -0 E-mail: mary.anne.bertola@saic.com Cara C. O Donnell Science Applications International Corporation 00 Georgetown Pike, F- McLean, VA 0 Phone: (0) - E-mail: odonnellcc@saic.com Jim Shurbutt, Ph.D. Federal Highway Administration 00 Georgetown Pike, T-0 McLean, VA 0 Phone: (0) -0 E-mail: jim.shurbutt@dot.gov WORD COUNT (0): Abstract () + Text (0) + Tables () + Figures ()

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of Evaluation of Truncated Arrow-per-Lane Guide Signs Bryan J. Katz, Erin E. Dagnall, Mary Anne Bertola, Cara C. O Donnell, and Jim Shurbutt 0 0 ABSTRACT Some freeway and expressway interchanges contain an interior option lane in which traffic in that lane can choose to exit or remain on the route at the split. The 00 MUTCD requires that either the Overhead Arrow-per-Lane or Diagrammatic guide sign designs be used for all multilane exits at major interchanges that have an optional exit lane that also carries the through route. However; some practitioners believe that it is cost-prohibitive and unnecessary to install Overhead Arrow-per-Lane guide signs at minor interchanges. Given the benefits of Overhead Arrow-per-Lane guide signs, a condensed version of an Overhead Arrow-per-Lane guide sign is needed for use where a full Overhead Arrow-per-Lane guide sign is not a reasonable option. A truncated version of the signs, which could be displayed on existing cantilever structures, would eliminate the expense of installing larger structures, while still reaping the benefits of the arrowper-lane sign concept. This study identified five signing alternatives for potential use at minor and intermediate interchanges with multi-lane exits and an option lane. Participants viewed videos of either Overhead Arrow-per-Lane guide signs, or of one of the four truncated alternatives. Participants were asked to indicate which lanes they could use to exit in some scenarios, and which to use to proceed through in other scenarios. Participants typically understood the alternatives that provided information about both through and exiting traffic (0% comprehension) whereas alternatives that provided information about exiting traffic, with little to no information about through traffic, resulted in relatively low comprehension (%).

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 0 0 INTRODUCTION Background Some freeway and expressway interchanges contain an interior option lane in which traffic in that lane can choose to exit or remain on the route at the split. The 00 MUTCD requires that either the Overhead Arrow-per-Lane or Diagrammatic guide sign designs be used for all multilane exits at major interchanges that have an optional exit lane that also carries the through route. The MUTCD advises that the Overhead Arrow-per-Lane guide sign design should also be considered for multi-lane exits with an option lane at intermediate and minor interchanges. However, some practitioners believe that it is cost-prohibitive and unnecessary to install Overhead Arrow-per-Lane guide signs at intermediate and minor interchanges. The MUTCD allows the use of a combination of conventional guide signing and regulatory lane-use signing when the inclusion of the Overhead Arrow-per-Lane guide sign design is not warranted. Some research has shown that conventional guide signs may produce fewer laneplacement errors and exit lane errors than certain types of diagrammatic signs. Conventional signs, however, lack the ability to display lane-choice configurations. In a survey, Brackett, Huchingson, Trout and Womack found that modified diagrammatic signs which provided separate arrows for each lane (such as the current Overhead Arrow-per-Lane sign design) were more effective in communicating lane assignment than the original diagrammatic sign. A recent study by Golembiewski and Katz (00) investigated the usage of an Overhead Arrow-per-Lane sign compared with original Diagrammatic guide signs. The results indicated that the Arrow-per- Lane sign type is appropriate for drivers of all ages, but is especially beneficial for older drivers. Therefore, the researchers concluded that, with regard to decision sight distance and correct lane choice, the Overhead Arrow-per-Lane sign provides superior navigation guidance. Given the benefits of Overhead Arrow-per-Lane guide signs, a condensed version is needed for use where a full Overhead Arrow-per-Lane guide sign is not a reasonable option. A truncated version of Overhead Arrow-per-Lane signs, which could be displayed on existing cantilever structures, would eliminate the expense of installing larger structures, while still reaping the benefits of the arrow-per-lane sign concept. The current research study was designed with the following two objectives in mind:. Identify truncated versions of the Overhead Arrow-per-Lane guide sign.. Evaluate effectiveness of the Overhead Arrow-per-Lane guide sign and any identified truncated versions compared to allowable MUTCD practices and other common practices. Stimuli Five signing alternatives were identified for potential use at minor and intermediate interchanges with multi-lane exits and an option lane. Each sign alternative is part of a three-sign series in which the first sign is located one mile from the exit gore, the second sign is placed approximately one half mile from the gore, and the third sign is placed near the gore. These alternatives are shown in Figure.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of Alternative : Truncated Arrow-per-Lane Guide Sign Series Alternative : Arrow-per-Lane Guide Sign Series Alternative : Truncated Arrow-per-Lane with Pull Through Series Alternative : Minor Interchange Guide Sign Series Alternative : Conventional Arrows Guide Sign Series 0 FIGURE Guide sign series alternatives tested. Alternative : Truncated Arrow-per-Lane Alternative is a truncated version of the Overhead Arrow-per-Lane guide sign design. This version includes only the up-arrows for the two right-most lanes (i.e., option lane and exit only lane), rather than also showing up-arrows for all adjacent through lanes. Alternative : Overhead Arrow-per-Lane Alternative is the Overhead Arrow-per-Lane guide sign design which is currently recommended through guidance statements for minor and intermediate interchanges that have multi-lane exits with an option lane (adapted from Figure E- in the MUTCD). This design includes one upward-pointing arrow per lane, which indicate to the driver the number of lanes and the direction of each lane.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 0 0 Alternative : Truncated Arrow-per-Lane with Pull Through Alternative is a truncated version of the Overhead Arrow-per-Lane guide sign design. This version includes the arrow per lane style signing for the two right-most lanes (i.e., option lane and exit only lane). Instead of including the route shield and cardinal direction for the through destination on the primary sign (as is done in alternative ), this information is presented on a secondary Pull-Through sign located on the same assembly. Both signs shown in Figure were used at each of the three locations in the sign series. Alternative : Minor Interchange Alternative is the current MUTCD recommendation for signing at intermediate and minor multi-lane exits with an option lane when Overhead Arrow-per-Lane or Diagrammatic guide signing is not warranted based on factors such as the extent of the need for the option lane and the type of the traffic exiting during high-volume periods (adapted from Figure E- in the MUTCD). This sign series is also used in conjunction with a post-mounted lane-use sign (R- Series). Alternative : Conventional Arrows Alternative is commonly used by many agencies to convey the presence of option lanes. METHOD The research was conducted in the Washington, D.C. metropolitan area. Participants were seated at a table in front of a laptop computer. Participants viewed videos of signs played on the laptop and were given worksheets for recording their responses. Each video displayed one of the five signing series alternatives. A researcher was seated next to the participant in order to run the videos and ensure that the participants were recording their responses properly. Each participant was exposed to only one of the five signing conditions. Prior to viewing the videos, participants were instructed that they were traveling to Interstate South toward Springfield, and were given instructions to indicate which lanes they could use to reach their destination, and which lanes they could not use to reach their destination. Each participant viewed two videos; one video had Interstate South/Springfield as the exiting destination, and the other video had it as the through destination. Interstate East toward Washington was always used as the alternate destination, i.e., the destination other than their target destination. The presentation order of the videos was randomized. The videos contained visible travel lanes, and ran at a simulated speed of mi/h. In each video, the lane point-of-view was from the left center lane. After passing the first sign in the series, the video was paused with the sign just out of view. Participants then marked on their response sheets which lanes they could and could not use to continue to Interstate. Lanes they could use were circled and lanes they could not use were marked with an X. The video continued in this manner, pausing after each sign in the series so the participant could record their response. Participants The study employed a sample of 00 research participants (0 male and 0 female). Males ranged in age from years to years (M =. years) and females ranged from years to 0 years (M =. years). Each participant possessed a valid state driver s license. Including instructions, signing of informed consent, and debriefing, the evaluation took about minutes to complete.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 RESULTS Error! Reference source not found. shows the percentage of participants who selected each lane at each sign in the series for exiting destination directions and for through destination directions. For example, Table indicates that, when considering the exiting conditions for participants who viewed Alternative, after viewing the first sign in the series % reported that they could use Lane to reach their destination, % reported they could use Lane to reach their destination, 0% reported they could use Lane to reach their destination, and % reported they could use Lane to reach their destination. Correct lane selections are highlight in red italics. As the first half of Table represents the exiting destination condition, only lanes and could have actually been used to reach their destination, thus they are highlighted in red; the second half of the table represents the through destination condition, thus lanes,, and could have been used to reach their destination. TABLE Percentage of Participant Selections for Each Lane at Each Sign in the Series for Exiting and Through Destination Directions Destination Exiting Through Sign Alternative Alternative Alternative Alternative Alternative Alternative Alternative Order in Series Lane Through Lane Through Lane Option Lane Exit st Sign.0 0.0.0 nd Sign 0.0 0 0.0.0 rd Sign 0.0.0 0.0 st Sign.. 0. 0. nd Sign.... rd Sign.. 0. 0. st Sign. 00.0 0.. nd Sign. 00.0 0.. rd Sign. 00.0. 0. st Sign.0 0.0 0.0 nd Sign 0.0 0.0 0 0.0 rd Sign.0.0 0.0 st Sign.... nd Sign.... rd Sign.... st Sign.0 0.0 0.0.0 nd Sign.0.0.0.0 rd Sign.0.0 0.0.0 Alternative st Sign 0.. 0..

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 Destination Sign Alternative Alternative Alternative Alternative Order in Series Lane Through Lane Through Lane Option nd Sign. 0. 0.. rd Sign 0. 0... Lane Exit st Sign.... nd Sign.... rd Sign. 0... st Sign 0.0 0.0 0.0.0 nd Sign.0.0.0 0.0 rd Sign.0.0.0 0.0 st Sign.... nd Sign.... rd Sign.... Data were considered as one set of responses per participant. Each set consisted of three questions, corresponding to each of the three signs in the series. If the participant selected all of the correct lanes, and only correct lanes, then he/she received a score of for the question. Conversely, if the participant did not select all of the correct lanes and/or selected one or more incorrect lanes, then he/she received a score of 0 for the question. Thus there were eight possible outcomes for each set of responses (0,0,0), (0,0,), (0,,0), (0,,,), (,0,0), (,,0), (,0,), and (,,) where the position in the set corresponded to the question following that sign number (e.g., a in the first position corresponded to answering correctly on the question following the first sign in the series). Generally speaking, the majority (%-%) of responses for each alternative and destination combination were either (0,0,0) or (,,). In other words, participant responses were usually either all incorrect (incorrect for all three signs in the series) or all correct (correct for all three signs in the series). However, this was not the case for Alternative for the exiting destinations. In this condition, only % of participants were either all correct or all incorrect, meaning that % of participants changed their responses at some point in the sign series. Three possible outcomes contained situations where the participant s responses were correct at some point during the sign series, but ultimately changed to incorrect by the end of the series; these outcomes were either (0,,0), (,0,0), or (,,0). These correct-to-incorrect changes constituted about % of the responses for the exiting destinations and about 0% of the responses for the through destinations. The remaining three possible outcomes contained situations where the participant s responses were incorrect at some point during the sign series, but ultimately changed to correct by the end of the sign series; these outcomes were either (0,0,), (0,,), or (,0,). Such incorrect-to-correct changes constituted about % of the responses for the exiting destination directions and about % of the responses for the through destination directions. Also, it should be noted that about % of such incorrect-to-correct changes occurred when using Alternative for the exiting destination directions. Additionally, of

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 0 all incorrect-to-correct changes when using Alternative for the exiting destination, approximately % of these changes came from the (0,,) outcome. Data were then analyzed using Generalized Estimating Equations (GEE), which is an appropriate methodology for categorical data when each participant contributes more than one measurement. A binomial response distribution and logit (log odds) link function were assumed. Variables included in the model were sign condition (Alternative, Alternative, Alternative, Alternative, or Alternative ), destination direction (exiting destination or through destination), sign view (after first sign in series, after second sign in series, or after third sign in series), and all corresponding second- and third-order interactions. Table presents the Type III statistics for this analysis. Condition was the only significant effect; χ () =., p < 0.0. TABLE Type III Statistics for GEE Model Analyzing Participant Responses by Condition, Direction, View, and Corresponding Interactions Df Chi-Square Pr > ChiSq Condition. <0.000 Destination 0.0 0. View 0. 0. Condition*Destination. 0.0 View*Destination.0 0. View*Condition. 0.0 View*Condition*Destination.0 0.0 Based on the information presented in Table, the order of responses in the set did not significantly affect how the participant would answer. Also, the majority of participant responses were either all correct or all incorrect for a given sign series. This means that regardless of how many times a participant viewed a sign in a series, his/her initial response after viewing the first sign typically did not change. However, as there were situations where a participant changed responses from incorrect to correct during the series, researchers decided to use the response on the last sign in the series to determine which sign alternatives were the best and whether participant performance was dependent upon the destination direction. For the remaining statistical analyses, overall correctness was defined by participant responses after viewing the third sign in the series. In other words, if the participant selected all (and only) the correct lanes that he/she could use to reach Interstate South toward Springfield after viewing the last sign in the series, then he/she was considered correct. The entire series of responses were not analyzed because, theoretically, as long as the driver understood which lanes they could use by the third sign, they should be able to make the appropriate lane maneuvers to reach their destination. Chi-squared tests were performed on contingency tables to examine the relationships between sign alternative, destination direction (exiting or through), and overall correctness. Likelihood ratio chi-squared statistics are reported.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 Overall Correctness by Sign Alternative A chi-squared test indicated that overall correctness and sign alternative are associated; χ () = 0., p < 0.00. The first five rows of Table show the percentage of correct responses by sign series alternative. Alternative, Alternative, and Alternative yielded more incorrect responses than correct responses. Alternative and Alternative yielded more correct responses than incorrect responses, with Alternative yielding the highest percentage of correct responses. When comparing Alternative and Alternative (the alternatives which resulted in the high percentages of correct responses), there was no significant difference in the proportion of correct responses (χ () =., p = 0.). Therefore, responses from these two sign alternatives were combined (M =.%) for comparison with the remaining sign alternatives. When comparing Alternative and Alternative (the alternatives which resulted in the two next highest percentages of correct responses, though both were below 0%), there was no significant difference in the proportion of correct responses (χ () = 0., p = 0.0). Therefore, responses from these two sign alternatives were combined (M =.%) for comparison with the remaining sign alternatives. When comparing the Alternative /Alternative combination with Alternative, there was no significant difference in the proportion of correct responses (χ () =., p = 0.). Therefore, the responses from these sign alternatives were also combined. The percentages of correct responses for the two resulting groups are shown in the last two rows of Table. TABLE Percentage of Overall Correctness by Sign Series Alternative (N = 0 for Each Sign Alternative) and by Combined Sign Series Alternatives Sign Alternative(s) Correct Responses Alternative 0.00 Alternative. Alternative. Alternative.0 Alternative. Alternative / Alternative Alternative / Alternative / Alternative..

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 When considering the comparison of Alternatives and combined (.%) against Alternatives,, and combined (.%), there was a significant difference in the percentage of correct responses; χ () =., p < 0.00. Overall Correctness by Destination Direction Results were analyzed to determine if destination direction (i.e., through or exit) was associated with correct responses. A chi-squared test indicated that overall correctness and destination direction are associated; χ () =., p < 0.0. The estimated odds of answering correctly was about % higher for exiting directions than for through directions. Table shows the percentage of correct responses by destination direction. TABLE Percentage of Overall Correctness by Destination Direction Destination Direction Correct Responses Exit. Through. 0 0 Overall Correctness by Sign alternative within Destination Direction Because destination direction had a significant effect on overall correctness, the sample was analyzed by sign series alternative within each direction. The first five rows of Table show the percentage of correct responses by alternative, within each destination direction. When considering only exiting destinations, overall correctness and sign alternative were significantly associated; χ () =., p = 0.00. Overall correctness and sign alternative were significantly associated when considering only through destinations as well; χ () =., p < 0.00. Alternative, Alternative, and Alternative yielded more correct responses than incorrect responses for exiting destinations. Similarly, Alternative and Alternative resulted in more correct responses than incorrect responses for through destinations; however, for through destinations, Alternative yielded more incorrect responses than correct responses. For both destinations, the highest percentage of correct responses occurred under Alternative and the lowest percentage of correct responses occurred under Alternative. When comparing Alternative and Alternative (these alternatives resulted in relatively high percentages of correct responses for each destination) there was no significant association between overall correctness and sign alternative within exiting destinations; there was also no significant association within through destinations. Therefore, the results from these alternatives were combined for further comparison to other sign alternatives within each respective destination. When comparing Alternative and Alternative (resulting in the next two highest percentages of correct responses for each destination) there was no significant association between overall correctness and sign alternative within exiting destinations; there was also no significant association within through destinations. Therefore, the results from these alternatives were combined for further comparison to other sign alternatives within each respective destination. When comparing the Alternative /Alternative combination with Alternative, there was no significant association between overall correctness and sign alternative within exiting

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page 0 of destinations; there was also no significant association within through destinations. Therefore, the responses from these sign alternatives were also combined. The percentages of correct responses for the two resulting groups are shown in the last two rows of Table. TABLE Percentage of Correct Responses by Sign Series Alternative and Combined Sign Series Alternatives within Each Destination Direction Correct Responses Sign Alternative Exiting Through Alternative.00.00 Alternative.. Alternative 0.. Alternative.00 0.00 Alternative.. 0 0 Alternative / Alternative Alternative / Alternative / Alternative.... Finally, when considering the comparison of Alternatives and combined against Alternatives,, and combined, there was a significant association between overall correctness and sign alternative within exiting destinations; χ () =.0, p < 0.00. The significant association was also present for through destinations; χ () =.0, p < 0.00. When considering the distribution of correct-to-incorrect responses within each destination direction, Alternatives and resulted in distributions of approximately / for each direction, whereas Alternatives,, and resulted in correct-to-incorrect response distributions of approximately 0/0 for exiting destinations and / for through destinations. For exiting destinations, the estimated odds of responding correctly were significantly higher for Alternatives and than for Alternatives,, and. For through destinations, the estimated odds of responding correctly were also significantly higher for Alternatives and than for Alternatives,, and. The extreme difference between the increases in odds is logical given the lack of similar response distributions between the two destination directions for Alternatives,, and. In other words, there is a larger increase in odds for the through destinations since more participants under Alternatives,, and responded incorrectly compared to the number of participants under Alternatives and who responded incorrectly.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 0 0 DISCUSSION AND CONCLUSIONS The results indicated that the majority of participants either responded always incorrectly or always correctly throughout the sign series. In other words, there was not much variation between signs, and in the series; most participants who started correct, remained correct throughout the series, and most that started incorrect, remained incorrect throughout the series. However, this was not the case for Alternative for the exiting destinations. In this condition, % of participants changed their responses at some point in the sign series. The majority of these responses (%) were incorrect-to-correct changes, meaning that the participant was incorrect at some point during the sign series, but ultimately changed to a correct response by the last sign in the series. One explanation of this may be the limited information provided by the first two signs in this series; these signs simply indicate Exit Only for Lane (right lane), and provide no additional guidance information regarding an option lane or the through destination; only a regulatory lane-use sign (R-) is used. Thus, when participants have the through destination, they may have seen the Exit Only information for Lane, and easily deduced that they could use the other three lanes to reach their destination. However it s not surprising that participants would be more likely to change their minds for the exiting destinations, as more interpretation of the Exit Only information may be required in order to respond. It is interesting to note that, when considering the exiting destinations for Alternative, approximately % of incorrect-tocorrect changes occurred after viewing the second guide sign. This implies that the R- regulatory lane-use sign may have given participants the additional information they needed to make a correct decision, as they were exposed to this sign after giving their response to the first overhead guide sign. The remaining % of incorrect-to-correct changes occurred after viewing the third guide sign in the series. This is also not surprising, as participants were exposed to another R- sign between the second and third guide signs, and were also given additional information on the exit lanes on the third guide sign. Because there was little variation between signs within a signing series for the majority of the alternatives, and because there were some participants who changed from incorrect to correct within a series, the remaining analyses looked at correctness based only on participant responses for the third sign in a signing series. It was found that the odds of understanding all lanes which can be used to reach a destination (when using either the exit or through direction) were significantly higher under Alternatives and than under Alternatives,, and. Alternatives and had comprehension rates of approximately % and %, respectively. In contrast, Alternatives,, and all had comprehension below 0%, with approximately 0%, %, and % correct responses, respectively. It is also important to consider sign alternative within destination direction. When participants were intending to exit, the odds of understanding all lanes which could be used to reach the destination were significantly higher for Alternatives and than for Alternatives,, and. When participants needed to use the through lanes, the odds of understanding all lanes which could be used were also significantly higher for Alternatives and than for Alternatives,, and. These results are not surprising as Alternatives and provide no information on the through destination. Although Alternative does provide some information on the through destination, participants may have misinterpreted the arrows indicating the option lane. Thus, without this additional information, it is possible that motorists may not understand if the sign is indicating that they should exit for South/Springfield, or that the entire road is South toward Springfield, except for the exit only lanes, which lead somewhere else.

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of 0 0 0 0 Of those participants who viewed Alternative, the percentage of correct responses for exiting destinations (%) and through destinations (%) indicates that participants do not understand all lanes which can be used for either destination direction. When considering lane selections for exiting destinations (see Table ) after viewing the third sign in Alternative, there was a % difference between the option lane (Lane ) and the adjacent correct lane (Lane ). This is a much greater difference than those seen for Alternative (0%), Alternative (.%), Alternative (%) and Alternative (.%). For through destinations, the difference between selections of the option lane (Lane ) and the adjacent correct lane (Lane ) for Alternative was not as high (%; see Table ), however it was still higher than the corresponding difference for most other alternatives. This indicates that, even though Alternative does provide information on the through destination, the concept of the option lane is not adequately conveyed by the signing alternative. It is important to recognize that participants in the Washington, DC area are familiar with Alternative as it is in common use in the area; however, participants were still unable to understand the intended meaning of the sign. Based on these results, Alternatives,, and are not recommended as suitable options for option lane signing, as they do not effectively convey all lanes which can be used when there are multi-lane exits with an option lane. Alternatives and both had high comprehension and were much more effective in conveying the intended information; this is presumably due to the fact that these alternatives provide the most information for each of the lanes. Therefore, of the alternatives tested, Alternative would be the best option for a truncated version of the Overhead Arrow-per-Lane guide sign at intermediate and minor interchanges which have multi-lane exits and an option lane. The use of pull through signs for Alternatives,, and may make a difference and should be considered in future research. ACKNOWLEDGEMENTS The authors would like to thank Dr. William Perez (SAIC) for his management oversight on the project. This research was funded by the Traffic Control Devices Pooled Fund Study (TCD PFS), which focuses on systematic evaluation of novel TCDs and employs a consistent process that addresses human factors and operations issues for each TCD idea. By pooling resources and expertise rather than to perform several independent research studies across the country, the PFS provides local and state agencies faster response to their needs and faster response to new technologies with the right assessment skills and tools to enable consistent TCD idea identification and evaluation. The PFS efforts address TCD issues identified by local and state jurisdictions, industry, and organizations and aid in the compliance to the MUTCD rule-making process and incorporation of novel TCDs into the MUTCD. REFERENCES. Manual on Uniform Traffic Control Devices for Streets and Highways. FHWA, U.S. Department of Transportation, Washington, D.C., 00.. Gordon, D.A. Evaluation of Diagrammatic Guide Signs. Highway Research Record,, pp. 0-,.. Brackett, Q., R.D. Huchingson, N.D. Trout, and K. Womack. Study of Urban Guide Sign Deficiencies. In Transportation Research Record: Journal of the Transportation Research

Katz, Dagnall, Bertola, O Donnell, and Shurbutt Page of Board, No.. Transportation Research Board of the National Academies, Washington, D.C.,.. Golembiewski, G. and B.J. Katz. Diagrammatic Freeway Guide Sign Design. Final Report. Traffic Control Device Pooled Fund Study, FHWA, U.S. Department of Transportation, 00.