Read a published paper of the study:Player preferences among new and old violins (screenshots of the article are attached below) and write a review that summarizes the statistical analysis used. Your paper must answer the following questions:

• What is the research question of the study?
• What method was used to collect data?
• What exploratory data analyses were used?
• What statistical inference methods were used?
• What is the conclusion of the study?
• What are the good features of the study?
• Is there any aspect that can be improved?

Screenshots of the article:

Player preferences among new and old violins Claudia Fritz .12, Joseph Curtin.', Jacques Poitevineau , Palmer Morrel-Samuels', and Fan-Chia Tao Lutheries-Acoustique-Musique, Institut Jean Le Rond d'Alembert, Unite Mixte de Recherche 7190, Centre National de la Recherche Scientificque, Universite Paris 06, 75015 Paris, France; Joseph Curtin Studios, Ann Arbor, MI 48103; "University of Michigan and Employee Motivation and Performance Assessment, Chelsea, MI 48118; and "D'Addario and Company, Farmingdale, NY 11735 Edited by Dale Purves, Duke University Medical Center, Durham, NC, and approved December 5, 2011 (received for review September 12, 2011) Most violinists believe that instruments by Stradivari and Guarneri absolutely." According to Langhoff (13), "any musician will tell "del Gesu" are tonally superior to other violins-and to new violins you immediately whether an instrument he is playing on is an in particular. Many mechanical and acoustical factors have been pro- antique instrument or a modern one." Neither of these hypo- posed to account for this superiority; however, the fundamental thetical statements has been tested, and, apart from recent pre- premise of tonal superiority has not yet been properly investigated. liminary results (14), the research literature contains no well- Player's judgments about a Stradivari's sound may be biased by the controlled studies on how violinists rate violins or whether they can violin's extraordinary monetary value and historical importance, but distinguish old Italian violins from old French or new American no studies designed to preclude such biasing factors have yet been violins by their playing qualities alone. published. We asked 21 experienced violinists to compare violins by In a recent wine-tasting experiment (15), subjects were given Stradivari and Guarneri del Gesu with high-quality new instruments. samples to taste while an MRI machine monitored brain activity. It The resulting preferences were based on the violinists' individual was found that increasing the stated price of a wine increased the experiences of playing the instruments under double-blind condi- level of "flavor pleasantness" reported by subjects; it also increased tions in a room with relatively dry acoustics. We found that () the activity in an area of the brain believed to encode for "experienced most-preferred violin was new; (i/) the least-preferred was by Stra- pleasantness." Could a violinist's preference for a Stradivari violin- divari; (iii) there was scant correlation between an instrument's age and, indeed, the pleasure he or she experiences in playing it-be in and monetary value and its perceived quality; and (iv) most players part attributable to an awareness of its multimillion-dollar price tag seemed unable to tell whether their most-preferred instrument was and historical importance, both of which may be signaled by its new or old. These results present a striking challenge to conven- distinctive appearance? Conversely, could the experience of playing tional wisdom. Differences in taste among individual players, along a new violin be negatively affected by the belief that it is still cen- with differences in playing qualities among individual instruments, turies from tonal maturity? To avoid any such biases, we tested appear more important than any general differences between new player preferences under double-blind conditions by using high- and old violins. Rather than searching for the "secret" of Stradivari, quality new violins together with distinguished "old Italians." future research might best focused on how violinists evaluate instruments, on which specific playing qualities are most important Materials and Methods to them, and on how these qualities relate to measurable attributes The experiment took advantage of the fine violinists, violin-makers, and violins of the instruments, whether old or new. gathered in September 2010 for the Eighth International Violin Competition of Indianapolis (IVCI), one of the most important international violin-playing tone quality | old Italian sound | subjective evaluation | music | perception competitions. Six instruments were assembled: three new and three old. The new violins (N1, N2, and N3) were each by a different maker and were between several days and several years old. They were chosen from a pool of violins A Imost all well-known violin soloists since the early 1800s have assembled by the authors, who then selected the three that they felt (i) had the chosen to play instruments by Antonio Stradivari or Giuseppe most impressive playing qualities and (ii) contrasted with each other in terms Guarneri "del Gesu," the two most celebrated craftsmen of the so- of character of sound. One was a Stradivari model; two were Guarneri models. called Golden Age of violin-making (ca. 1550 to ca. 1750). A long- The old violins consisted of one by Guarneri del Gesu (ca. 1740) and two by standing goal of violin research has been to correlate the playing Antonio Stradivari (ca. 1700 and ca. 1715). These violins were loaned with the qualities of these instruments with specific attributes of their stipulation that they remain in the condition in which we received them physical structure and dynamic behavior, and yet "no [objectively (precluding any tonal adjustments or even changing the strings) and that their measurable] specification which successfully defines even coarse identities remain confidential (hence the very general descriptions that fol low). The earlier Stradivari (01) was once the principal instrument of a well- divisions in instrument quality is known" (author's italics) (1). Many known 20th century violinist and currently belongs to an institution that loans factors have been proposed and/or investigated to account for the it to gifted violinists. It came to us from a soloist who had used it for numerous presumed tonal superiority of old Italian violins, including prop- concerts and several commercial recordings in recent years. The later Stradivari erties of the varnish (2, 3), effects of the Little Ice Age on violin (03) is from the maker's "Golden Period" and has been used by a number of wood (4), differences in the relative densities of early- and late- well-known violinists for concerts and recordings. The Guarneri del Gesu (02) is growth layers in wood (5), chemical treatments of the wood (6, 7), from the maker's late period, during which he made some of his most cele- plate-tuning methods (8), and the spectral balance of the radiated brated violins. The combined value of the old violins is approximately $10 sound (9-11). However, although correlations between violin million-roughly 100 times that of the new violins. Although the instruments acoustics and perception have been attempted (12), the funda- were not all set up with the same strings, all had the very typical combination of a steel E with metal-wound, synthetic-core strings for the rest. All strings mental premise of tonal superiority has not yet been properly in appeared to be in good condition. vestigated. Stradivari and Guarneri del Gesu may well be the greatest violin makers ever, but it takes an expert opinion based on visual and historical (rather than tonal) evidence to say whether a Author contributions: C.F., J.C., P.M.-S., and F.-C.T. designed research; C.F., J.C., and F.-C.T. particular example is genuine. Playing and listening tests never performed research; C.F., J.C., and J.P. analyzed data; and C.F., J.C., and J.P. wrote enter the authentication process, suggesting the difficulty of re- the paper. liably rating playing qualities-and that they may not correlate well The authors declare no conflict of interest. with an instrument's age and maker. This article is a PNAS Direct Submission. Weinreich (1) argues that any experienced player can classify C.F. and J.C. contributed equally to this work. a violin as a "student," "decent professional," or "fine solo" in- 2To whom correspondence should be addressed. E-mail: claudia.fritz@upmc.fr. strument; furthermore, "the judgment would not take more than This article contains supporting information online at www.pnas.org/lookup/supplidoi:10. about 30 s, and the opinions of different violinists would coincide 1073/pnas. 11 14999109/-/DCSupplemental.Numbers of subjects and instruments were small because it is difficult to persua de the owners of fragile, enormously valuable old violins to relea se th em for extended periods into the hands of blindfolded strangers. Many of the 21 subjects were involved with the IVCI as contestants (four), jury members (M0), or members of the Indianapolis Symphony. Nineteen subjects described themselves as professionals, 10 had advanced degrees in music, and 2 were later chosen as competition laureates. The subjects ranged in age from 20 to 65 y, had played violin for 15-61 y, and owned violins between 3 and 323 y old, with approximate values ranging from $1,800 (US) to $10 million {Si Text). Although we believe all subjects were sufficiently skilled for their preferences to be meaningful, we are aware that players with different levels of expertise may form their preferences on different grounds. This factor, however, is outside the scope of our study. To attract participants, potential subjects were told that they would have the chance to play a number of fine violins, including at least one by Stradivari. No other information about test instruments, including the number involved, or whether they were old or new, was disclosed. Subjects we re scheduled for individual, 'I-h sessions, before which they were given instructions to read {Si Text) and a questionnaire and consent form to fill out. When trying out instruments, violinists typically use their own bows, which through constant use have become, in effect, extensions of their bow arms. In light of this practice, we asked subjects to bring their own bows. For the four who did not, a single good-quality bow was provided. Most violinists prefer to try out violins in a room with relatively dry acoustics, where the direct sound from the instrument is not so much colored by room reflections. Sessions were there- fore conducted in a hotel room whose acoustits seemed well-suited to the task. We are aware that room acoustics may influence a player's preference for one instrument or anoaer. However, that is a separate question not covered in this study. Throughout the sessions, subjects wore modified welders' goggles, which, together with much-reduced ambient lighting, made it impossible to identify instruments by eye. To mask any distinctive smells, a dab of scent was put underthe chinrest of each violin.The hotel room was divided into two are as by a cloth screen. To preserve double-blind conditions, violins were passed from behind the screen to a researcher wearing goggles, who laid them on a bed in the order received. This study explores player preferences under two sets of conditions. One set, designed to maximize ecological validity, emulated the way players choose instruments at a violin shop, where they typically try a selection of instruments before selecting one to take home for further testing. All six test instruments were laid out in random order on the bed. Subjects were then given 20 min to choose (i) the single instrument they would "most like to take home with them" and {ii} the instruments they considered "best\" and \"worst" in each of four categories: range of tone colors, projection, play- ability. and response. These terms, all commonly used by players when evaluating instruments, were left undefined. If a term lacked clear meaning for a subject, heishe was told not to choose in that category. Although projection can, by definition, be judged only at a distance by a listener, players regularly estimate projection when testing a violin. They typically acknowledge (as did many of our subjects) the provisional nature of such estimates and the need to retest in a large hall with trusted listeners. Note, however, that our experiment was designed to test not the objective qual- ities of the instruments but rather the subjective preferences of the subjects under a specific set of conditions. When making the bestiworst selections, equal ranking between instruments was permitted {i.e., several could tie for best or worst), as was refraining from choosing. Subjects were free to play the instruments in any order and in any manner they saw fit, including switching back and forth among them. They were also encouraged to comment out loud about the instruments and se- lection process. A researcher made notes of the subjects' comments but responded to them only to confirm what had been said. At the end of the session, subjects were invited to guess the "making-sd'tool" of their take-home instrumentsan indirect way of assessing their ability to distinguish new instruments from old ones. Our second set of test conditions, designed with the statements of Wein- re ich and Langhoft in mind, asked subjectsto assess instruments ratherquickly. Each subject was presented with a series of 10 pairs of violins. For each pair, subjects were given 1 min to play whatever they liked on the first violin, then another minute for the second violin, without switching back and forth be- tween them. The minute began with the first played note, including any tuning, and ended with the ringing of a bell. Subjects were then asked to state which violin they preferred. Unbeknownst to them. each pair consisted of a new and an old violin. Our set of three and three thus allowed for nine possible pairings. The order of the pairsand of the instruments within each pairwas randomized to avoid presentation order effects. As a rudimentary test for consistency, one of the nine pairs was presented twice. The retested pair was positioned randomly but with at least one other pair-wise compari- son separating test from retest. The pair-wise comparisons were conducted at the beginning of each session and will thus be referred to as \"part 1" of the experiment and the take-homei'bestiworst selections as "part 2." We believed that (i) part 1 should not be conducted after the subjects had played the violins for 20 min and (it) part 1 was not likely to affect part 2 judgments because subjects were given no information about any (possible) relationships between the violins in parts 'I and 2. Results and Discussion When analyzing player preferences in part 1, we omitted the ret ests and considered only the primary nine pairs, where each in strument was played just once by each subject (SI Text). From these nine pairs, the mean number of times an old violin was chosen was 3.7. The twosided 95% condence interval (CI; all CIs are two-sided 95% intervals through the article) is [2.8; 4.5]. Al- though this interval leaves room for old and new to be equally preferred, equality is in itself a radical notion given prevailing opinions about old violins. Table 1 shows the number of times each violin was chosen in each of the nine newrold pairings. In the six pairings not involving 01, the other ve violins were chosen about equally often. By contrast, whenever 01 was paired with a new violin, it was chosen markedly less often. It seems that under these test conditions, only a conspicuously leastpreferred violin differentiates itself. That violin happened to be a Stradivari (m. 1700), and its consistent rejection appears to drive the overall preference for new violins seen above. We found no evidence that this preference was af fected by the age of the subjects' own violin {SI Text). Considering now the retested pairs, just 11 of 21 subjects (52%) made the same choice twice. The CI is [30%; 74%], meaning no rm conclusions about player consistency can be drawn. Note, however, that if subjects perform no better than chance in such a test, two possible conclusions might be drawn: (1' ) the instruments are about equal in overall quality (as suggested in Table 1), which means that forcing subjects to choose among them (in effect) forces random choices, where consistency cannot be expected, and (if) subjects cannot choose consistently under part 1 conditions, which may therefore be unsuited to studying player preferences {SI Text). In part 2, subjects were free to play any violin against any other, new or old, and to divide time between the instruments as they saw t. Fig. 1 shows how often each violin was chosen as takehome choice (dark gray bar) and then as best or worst in four categories. Eight subjects voluntarily identied their least favorite instru ments; these are shown in black beneath the take-home bar. Eight subjects had difculty deciding which of two violins to take home: the times a violin was a close second is shown above the takehome bar in lighter gray. In contrast to part 1, where ve violins were chosen about equally, the violins now differentiate themselves more clearly. A single new instrument, N2, stands out as the most preferred: it was Table 1. Number of times each violin was chosen in each of the newrold pairings Pairs New violin chosen Old violin chosen CI for old, % N1-01 15 6 [11; 52] N2-01 18 3 [3; 36] N3-01 15 5 is; 4?] N1-02 10 11 [30; T4] N2-02 10 11 [30; 1'4] N3-02 11 10 [26; T0] N1-03 11 10 [26; 70] N2-03 11 10 [26; T0] N3-o3 10 11 [30; T4] PSYGIOLOGICAL AND COGNITIVE 5C IENES Close 2nd Tone colors Play-ability 13 Take-home Response Projection ' Best 10 Least favorite 3 3" 6 5 5 5 5 1 1 :1. 3 3 3 2 2 2 1 1 1| 1 l l 'l l l 1' 2J 2 2 3 3 3 3 3 i 5 5 3" N1 N2 N3 01. O2 03 Fig. 1. second and least-favorite {above and below take-home block, respectively}. chosen eight times as takehome, three times as close second, never as leastfavorite, and just three times as worstinacategory. By contrast, 01 (ca. 1700 Stradivari) was chosen once as take home, once as close second, six times as leastfavorite, and 16 times as worstinacategory. Although each violin was the takehome choice of at least one subject, four violins were also the leastfavorite for at least one subject. This wide divergence in individual taste carries through into the four categories: With the sole exception of N2's pro jection, each instrument was chosen as best and worst at least once in each category. Unsurprisingly, each subject rated their take home violin as best in at least one category (SI Text). Just 8 of 21 subjects (38%) chose an old violin to take home. Given the small sample size, this disinclination toward the old cannot be condently inferred to experienced violinists in general (CI [18%; 6292]). Still, the upper limit for the CI is not high; moreover, the fact that a new violin was chosen over examples by Stradivari and Guameri del Gesu by 13 experienced violinists [including both jury members, who compared N2 and N3 favorably with their own Stradivari and Guarneri del Gesu violins; SI Terr) stands as a bracing counterexample to conventional wisdom. How consistent were the subjects? Of the 15 who chose new violins more often than old ones in part 1, 7 later chose old violins to take home. Against this, ve subjects who chose old violins more often in part 1 later chose new violins to take home (SI Text). By this measure, just 9 of 21 were consistentalthough this nding seems unsurprising given the way preferences shifted as time was spent with individual instruments (Sir Tart). What was consistent through parts 1 and 2 was a preference for new violins and a specic dislike for 01. Can violinists tell new violins from old by their playing qualities alone? In coding the bestlworst selections in the four categories, violins were given a score of +1 for \"best\" in a category, 1 for \"worst,\" and l] for neither \"best\" nor \"worst\". This coding allowed us to accommodate subjects who selected as many as six violins as \"best\" or \"worst\" (e.g., by saying \"all are equally good\") or as few as none (e.g., by saying \"none are bad\"). Because the scores range from 1 to +1, a difference of 0.50 is a huge effect and one of 0.33 is quite strong. Results are presented in Fig. 2. {Strad c.1i'00) {Guarneri'delC-iesu'l {Stradc.1?15} Number of times each violin was selected as take-home and then as best or worst in four categories. Also shown are volunteered selections for close Subjects rated new violins signicantly more highly (P