Experimental Design. The Contractor shall select a rigorous experimental design that controls for confounding factors and other threats to the validity of the evaluation. The Contractor shall discuss the strengths and limitations of the chosen experimental design, including how it accounts for the effects of confounding factors in isolating impacts of the deployment. If using control/treatment groups, methods for recruiting participants of these groups, identifying the right type of participants, determining the appropriate sample size, and potential issues associated with sample size shall be discussed. Note that recruiting of participants is required only if the sites’ provided data and experimental designs are not sufficient to control for internal and external threats to validity of evaluation.
Experimental Design. The Contractor shall select a rigorous experimental design that controls for confounding factors and other threats to the validity of the evaluation. The Contractor shall discuss the strengths and limitations of the chosen experimental design, including how it accounts for the effects of confounding factors in isolating impacts of the deployment. If using control/treatment groups, methods for recruiting participants of these groups, identifying the right type of participants, determining the appropriate sample size, and potential issues associated with sample size shall be discussed. NOTE: The recruiting of participants is required only if the sites’ provided data and experimental designs are not sufficient to control for internal and external threats to validity of evaluation. The Contractor shall refine the MEP evaluation plans and submit for review to the COR and other USDOT representatives, and revise based on comments received. The Contractor shall submit to the COR, the Revised Plans and corresponding Comment Resolution Reports that document how each substantive comment was resolved. The COR, in consultation with the Contractor shall determine if a meeting is necessary to discuss any of the comments (depending on the nature and quantity of the comments). Upon the COR’s approval of the Comment Resolution Reports, the Contractor shall make revisions (as per the Reports) and submit final versions of the MEP evaluation plans.
Experimental Design. We initially considered four treatments, with each of the four types of game presented in Table 1 corresponding to a treatment. Following the advice of a referee we subsequently considered a further two treatments with variations on the vector and full-agreement game (more details to follow shortly). We reiterate that our standard treatment corresponds to the benchmark 14 In a symmetric game ‘split the cost proportionally’ and other alternatives are equivalent to split the cost equally. 15 This more liberal interpretation of label scrambling would make no difference in a standard game or vector game. treatment used in the threshold public goods literature. The game used in the standard treatment with feedback, vector treatments, and full agreement treatments differ as detailed in Section 2. Each experimental session was divided into three parts, as summarised in Table 4. In part 1, subjects played a game with parameters corresponding to those in the symmetric game, as already detailed in Table 2, for 10 rounds. In part 2 they played a game with parameters corresponding to those in the asymmetric game for a further 10 rounds, and in part 3 they played a game with parameters corresponding to those in the very asymmetric game for a final 10 rounds. The type of game played, standard, standard with feedback, vector or full agreement, was the same in all three parts of a session. Note that subjects retained their role within the group throughout a part. Thus, a subject endowed with, say, 70 in an asymmetric game was endowed with 70 in all 10 rounds. The groups, of five, were randomly assigned at the beginning of each part but remained fixed during the part. Fixed matching during each part of the session allows us to look for dynamic and learning effects as observed in previous threshold public good experiments (e.g. Cadsby et al. 2008). Indeed, given our interpretation of the vector of contributions as a form of indirect communication it is natural to think of the 10 rounds within each part as part of one big game. With this interpretation the final round of the ten takes on special importance as culmination of the game. In this last round there is nothing to be gained by indirect communication and so the only relevant objective is to maximize round payoff. The use of three different sets of parameters allows us to consider symmetric and asymmetric games.16 More specifically, the use of the benchmark parameters in part 1 allows an unambiguous comparison of behav...
Experimental Design. Using a within-subjects design, three different image comparison presentation styles were manipulated, side-by side, automatic flicker and manual flicker - where the participant used an on-screen button to change between each image. Three separate classification interfaces that varied in relation to these presentation styles were employed, again in conjunction with a questionnaire including NASA Task Load Index (TLX) type statements to assess volunteer opinion and perceived workload. The questionnaire also allowed ‘free-text’ responses so participants could raise issues and add context to their responses.
Experimental Design. We initially considered four treatments, with each of the four types of game presented in Table 1 corresponding to a treatment. Following the advice of a referee we subsequently considered a further two treatments with variations on the vector and full-agreement game (more details to follow shortly). We reiterate that our standard treatment corresponds to the benchmark 14 Iﾐ ; ゲ┞ﾏﾏWデ�ｷI ｪ;ﾏW けゲヮﾉｷデ デｴW Iﾗゲデ ヮ�ﾗヮﾗ�デｷﾗﾐ;ﾉﾉ┞げ ;ﾐS ﾗデｴW� ;ﾉデW�ﾐ;デｷ┗Wゲ ;�W Wケ┌ｷ┗;ﾉWﾐデ デﾗ ゲヮﾉｷデ デｴW Iﾗゲデ equally. 15 This more liberal interpretation of label scrambling would make no difference in a standard game or vector game. treatment used in the threshold public goods literature. The game used in the standard treatment with feedback, vector treatments, and full agreement treatments differ as detailed in Section 2. Each experimental session was divided into three parts, as summarised in Table 4. In part 1, subjects played a game with parameters corresponding to those in the symmetric game, as already detailed in Table 2, for 10 rounds. In part 2 they played a game with parameters corresponding to those in the asymmetric game for a further 10 rounds, and in part 3 they played a game with parameters corresponding to those in the very asymmetric game for a final 10 rounds. The type of game played, standard, standard with feedback, vector or full agreement, was the same in all three parts of a session. Note that subjects retained their role within the group throughout a part. Thus, a subject endowed with, say, 70 in an asymmetric game was endowed with 70 in all 10 rounds. The groups, of five, were randomly assigned at the beginning of each part but remained fixed during the part. Fixed matching during each part of the session allows us to look for dynamic and learning effects as observed in previous threshold public good experiments (e.g. Cadsby et al. 2008). Indeed, given our interpretation of the vector of contributions as a form of indirect communication it is natural to think of the 10 rounds within each part as part of one big game. With this interpretation the final round of the ten takes on special importance as culmination of the game. In this last round there is nothing to be gained by indirect communication and so the only relevant objective is to maximize round payoff. The use of three different sets of parameters allows us to consider symmetric and asymmetric games.16 More specifically, the use of the benchmark parameters in part 1 allows an unambiguous comparison of behav...
Experimental Design. Each experimental session was comprised of four parts followed by a survey: an initial trust game (a.k.a. investment game) in which subjects played both the part of the trustor and the trustee (Part 1); 10 rounds (“days”) of trading in the market game under no institutions (Part 2); 10 rounds of trading under either partial enforcement system (PES) or impartial enforcement system (IES) (Part 4); a final trust game (Part 3), for a total of 24 decisions per subject.
Experimental Design. Operational “boundaries” of PSTA systems are being investigated in mesocosms at an experimental facility near the outflow of STA-1W. Triplicate flow ways with a local limerock substrate were established under each of four water depth treatments. The first two treatments are static in depth. Shallow treatments (23 cm) and deeper treatments (46 cm) consist of 4 tanks (each 1.8 m2) plumbed in series. These tanks were initially established in September 2013, under constant flows that provide a hydraulic retention time (HRT) of 5 and 10 days for the shallow (23 cm) and deep (46 cm) flow ways, respectively. Delivery of a constant flow rate to both shallow and deep tanks insures equal P mass loading rate (PLR) to those treatments on an area basis. In January 2014, additional mesocosms were established to test PSTA performance at greater water depths. Six new flow ways were constructed using larger tanks (2.8 m2 per tank) plumbed two in series (Figure 1 and Figure 2). These systems were initially established at 46 cm depth, and flows are being delivered to provide equivalent HRT and PLR conditions to the existing mesocosms operating with 4 tanks-in-series at 46 cm depth. The first tanks in series of the new flow ways receive an equivalent PLR to the first half (first 2 tanks) of the 4-in-series systems. This approach enables a comparison of “midpoint” and “outflow” positions with equivalent HLR and P loading across static and variable-depth treatments. Key operational parameters of these systems are outlined in Table 1.
Experimental Design. As in the first study, we conducted a between-subjects deception experiment, but with only two conditions: a textured agreement and a plain-text control condition. Subjects were asked to down- load, install, and use a single image manipulation application (as opposed to three in the first experiment). The same instrumented installation environment was used as before. However, the dis- tractor task of using the application was not actually performed by participants (though the instructions asked them to use the soft- ware after installing it to rate its usability). Instead, participants were interrupted after reaching the point in the software installa- tion process where the software would actually be installed. Instead of installing the software, the participant was stopped and given a content quiz to test how much information they absorbed from the agreement process. This approach minimized the time between exposure to the agreements and taking the quiz. Scroll position (%) Scroll position (%) 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 100 80 60 40 20 0 0 50 100 150 0 50 100 150 6.2 Procedure 100 80 60 40 20 0 0 50 100 150 0 50 100 150 Scroll position (%) Scroll position (%) Scroll position (%) As before, subjects were given a written scenario and instructions after obtaining verbal consent. The scenario indicated that they had recently received a digital camera, but lacked software to perform basic manipulations of the images. Accordingly, they were told to imagine they had just found the website of an image manipulation application. The instructions asked them to download, install, and evaluate the application, and to decide whether they would con- tinue to use this program on their home computer. Once they had reached a decision, the instructions indicated that they would be given a questionnaire. After receiving and reading the instructions, subjects had the op- portunity to ask questions. They were then seated at a desktop com- puter with a web browser already opened to the download page of the application, called “Program A”. Participants were then able to download and run the installer. After clicking the “Next” button on the software agreement screen in the installer, a full-page screen in- formed participants that the task portion of the study was complete, and that the researcher will set up the questionnaire ...
Experimental Design. The experimental design is summarized in table 1 below. Table 1. Experimental design Animals Rats WISTAR RjHan:WI, male Nanomaterials TiO2 NM100 and NM101, CeO2 NM212 + TiO2 NM105 as reference NM Exposure method Unique instillation after hyperventilation Theoretical dose 500, 50, 5 μg/animal (0.125, 0.0125, 0.00125 μg/cm2 *) Exposure duration 3h, 24h, 5d, + (35d and 90d for biodistribution and histology) Endpoints On bronchoalveolar Lavage fluids(BALF): Cytotoxicity, Inflammation, Oxidative stress (3h, 24h, 5d) On blood smear: μ-nucleus assay (5d group only) Biodistribution: lungs, tracheobronchial nodes, spleen, liver, kidneys (35 and 90d) On animal tissue: Histology (35d and 90d only) Characterization of suspensions Dynamic Light Scattering (DLS) Dosimetry ICP-MS * assuming 4000 cm2 for total rat lung alveolar surface [1]
Experimental Design. A detailed description of the experimental conditions and an illustration of the set up can be found at 2.2.1. Briefly, physiologically relevant concentrations of the three major water soluble antioxidants associated with the RTLF were prepared as described in 2.1.3.2. Exposure of synthetic RTLF was carried out in a 5.6 litre purpose-built Perspex chamber, flushed with NO2 or O3 at 50, 150, 400 or 1000ppb, using the general conditions described in section 2.2.1. At 30, 60, 90 and 120 minute intervals throughout the exposure, the chamber was opened very briefly, samples removed and stored at - 80°C for up to a week prior to analyses described in 2.1.4 and 2.1.5.
