Intro



Experiments with Set

Two studies on Set game play

Johanna M. B. Craig

December 15, 2004

[pic]

Abstract 3

Introduction 3

Background 5

The Game of Set 5

Purpose 7

Research Setting and Population 10

Experiment 1 – The Training Study 13

Method 13

Participants 13

Materials 13

Design 13

Procedure 14

Results 14

Discussion 15

Experiment 2 – The Think-aloud Study 17

Method 17

Participants 17

Materials, Design and Procedure 17

Results 18

The Novice Player 18

The Expert Player 19

Discussion 20

Summary and Conclusions 21

References 22

Appendix A – Participant Questionnaire – November 2004 23

Appendix B – Stroop Tasks 24

Appendix C – Remote Associates Pre-test 24

Appendix D – Remote Associates Post-test 24

Appendix E – Set Game Resources 25

Appendix F – Training Study Results 26

Abstract

Set is a card game often touted to help individuals become better critical thinkers and problem solvers, yet no previous empirical research has been done on the cognitive aspects of its game play. In this paper I describe two studies of the cognitive aspects of Set. The first experiment is a training study, with the experimental group training on Set game play. Pre and post-test measures of inhibition/interference, convergent thinking, and Set game play ability are discussed. The second experiment is a novice-expert think-aloud experiment. Understanding expertise is relevant and important as it can offer insight into the thinking and problem solving strategies employed by Set players. I conclude with some thoughts on how to improve the first experiment as well as some ideas for further study.

Introduction

Games are often used in both formal and informal learning environments as a fun and challenging extension to typical learning activities. Games like Candy Land, Go Fish! or Chutes and Ladders are often played with preschool-aged children to help them learn their colors and develop “number sense.” As children mature, the educational game palette expands to include card games, strategy games, word games, and trivia games as well as a wide variety of computer “edutainment” titles.

One such title is the card game Set, which is used in a variety of educational settings. Parents play Set with their kids at home, teenagers play Set with their friends during free periods at school, and teachers often use Set with students in both gifted and special education programs. There is a small amount of research on the math of Set (Van Brink, 1997; Quinn, Weening, and Koca, Jr., 1999) but no one has ever really investigated the cognitive basis for the game. It is clear that Set game play involves pattern recognition and working memory to some extent, but why would it make kids better problem solvers and increase higher-order thinking skills?

In search for the answer to this question, I designed two experiments in hopes of acquiring a greater understanding of the cognitive dimension of Set game play. The first experiment addresses the mechanism of Set game play and posits that training on Set game play could foster improvement in specific measures of cognitive functions. The second experiment addresses the novice-expert paradigm with regard to Set game play strategies. To this end, I developed the following two research questions.

1) What does playing the game of Set cognitively entail?

2) What do novice and expert Set strategies look like?

In this paper I will describe the background of my study, including the game play of Set, the purpose of my two experiments, the research setting, and the participants’ backgrounds. I will then discuss the methods and results of each of my experiments. I finish with a discussion of my findings and based on these findings, make some suggestions for further research on this topic.

Background

The Game of Set

A Set deck consists of 81 cards with three variations across four attributes: number, color, symbol, and shading (see Figure 1). No two cards are alike.

a) Number – one, two, three

b) Color – red, purple, green

c) Symbol – oval, squiggle, diamond

d) Shading – solid, stripes or open

Figure 1: Attributes of a Set card

[pic]

A 'Set' consists of three cards in which each feature is either the same on each card or is different on each card. That is to say, any feature in the set of three cards is either common to all three cards or is different on each card. Figure 2 is a set because they are the same number (two), the same color (red), the same shape (ovals) and they are all of different shading. Figure 3 is an example of the “nothing in common set” as no card has anything in common with the other 2 cards in the set. They are different in number, color, symbol and shading.

Figure 2: A typical set

[pic]

Figure 3: The “nothing in common” set

[pic]

The game begins with the first 12 cards drawn from the deck arranged in a 3 x 4 rectangle. In the solitaire version, the player makes a set, puts the 3 cards aside and deals 3 more. In the group version, players call out the word “set” when they see one and are required to show (and get approval from) the other players before they can pick up their set and 3 new cards are dealt from the deck. If the players agree that no set can be found, 3 additional cards are put into play temporarily. Once a set is found they are not replaced and the board returns to 12. Game play ends, in both the solitaire and group version, when all the cards are dealt and no further sets can be found. In the group version, the player with the most sets is the winner.

There is also a computer-based version of the game Set, known as the “daily puzzle.” It is available at and is played slightly differently from the traditional card game of Set. Here the player, working alone or in a small group, must identify the six sets embedded in the 12 card display. In this case, the cards are re-used and the set is not “collected” – it is merely displayed on a scoreboard so the player can keep track of how many and which kinds of sets she has already made. Game play ends when the player finds all 6 sets or gives up. There is a new puzzle every day.

Purpose

The purpose of these two studies is to gain insight and understanding as to the implications for learning embodied in the game of Set. The goal of the first study, known as the training study, is to discover the mechanisms invoked by playing Set and identify the executive functions with which game play correlates. Secondarily, the goal is to identify whether training on Set game play might yield a measurable change in performance of some common tasks like the Stroop Color/Word Test (Stroop, 1935, as referenced in Dempster and Corkill, 1999) and the Remote Associates Test (Mednick and Mednick, 1967, as referenced in Leahy and Harris, 2001), or Set game play itself.

One possibility for the mechanism of Set is the interaction of interference and inhibition. According to Dempster and Corkill (1999), interference is classically defined as a decline in task performance caused by irrelevant information or behavior. Inhibition is then defined as the “active dampening process” (p. 2) by which the interfering information is suppressed. Current work on interference and inhibition does not always make a clear distinction between these two concepts. While they are not the same construct, they are closely related both empirically and theoretically. Set may work through these mechanisms, as a player must suppress information presented on some cards to attend to information on the cards that compose the prospective set.

The game of Set might also work through the process of convergent thinking, where a person begins with highly disparate, weakly-associated elements and finds some connection between them (Mednick, 1962). This is traditionally a measure of one type of creativity, but may very well tap into the same kind of skill required to be a successful Set player.

The executive control of task switching (Miyake et al, 2000) may be another function employed in the game of Set. Also known as attention switching, this is the function that allows a person to shift back and forth between mental sets, tasks, or operations. Each attribute of a Set card could be considered a different mental set. In order to identify a set, a person must shift back and forth through the variety of mental sets that comprise the various card attributes. In this way, the act of generating a set could be conceptualized as a series of attentional shifts.

While the specific mechanism(s) of Set are still unknown, there is a large body of research on the various executive functions. Working memory capacity, for instance, has often been regarded as both fixed and an indicator of general cognitive functions (Engle, Kane, and Tuholski, 1999). The findings of Klingberg, Forssberg and Westerberg (2002) seem to contradict this notion. These authors have recently shown that, through intensive computer training, working memory capacity can in fact be improved. Klingberg et al. were able to show effects with 7 ADHD children as well as 4 healthy male adults. Not only did the subjects in the treatment group show improvement in the particular training tasks, but they also achieved a higher score on the Ravens Progressive Matrices, a task for which there was no training. Another experiment by Olesen, Westerberg, and Klingberg (2004) found that working memory training produced a significant change in brain activity as measured by a functional magnetic resonance imaging study. These findings may be evidence of “…training-induced plasticity in the neural systems that underlie working memory” (Olesen, Westerberg, and Klingberg, 2004, p. 1). These two studies are the basis for the training aspect of this Set study.

The goal of the second study, known as the think-aloud study, is to gain an understanding as to how novice and expert strategies differ in Set game play. Understanding expertise is relevant and important as it can offer insight into the thinking and problem solving strategies employed by Set players. Experts possess abilities beyond just general memory or intelligence and tend to see problems in a very different way than even experienced novices do (Bransford, Brown, and Cocking, 2000). Expertise has been studied in wide variety of domains including chess, medical diagnosis, physics, and history. In each of these cases, expertise in a domain helps people develop a sensitivity to patterns of meaningful information that are not available to novices (Bransford, Brown, and Cocking, 2000). Though Set is a very small domain when compared to a game like chess, there are still a number of strategies employed in game play. As such, the think-aloud study is based on de Groot’s (1965) idea of think-aloud protocols for capturing expertise in chess (Bransford, Brown, and Cocking, 2000).

Research Setting and Population

This research took place in conjunction with an upper level Psychology seminar conducted by a faculty member who is jointly appointed to the Psychology department and to the Combined Program in Education and Psychology at a large Midwestern research university. The 10 subjects were between the ages 14 and 36. The mean age was 26.8. There were six males and four females. One subject was in high school, one was an undergraduate, five subjects had completed bachelor’s degrees, and 3 subjects had attended some college. Six of the subjects had been part of a “gifted and talented” program in elementary school, and 5 of these programs had a “puzzle or logic problem” component. One subject, though not in a specific program, spent time in elementary school working on logic problems and puzzles. Four of the subjects had played Set before this study.

The study participants had a wide variety of favorite school subjects (see Figure 4), with English/Literature being the most common, followed by physical science. There was less variation in favorite type of board game. Strategy games like Chess, Risk, and Mastermind were significantly favored, followed by word games like Scrabble and Boggle (see Figure 5).

Finally, there was again a fair amount of variation among subjects in regard to favorite computer games (see Figure 6). Simulation games like SimCity or the Sims and strategy games like Starcraft or Age of Empires were the favorites, though not by an overwhelming majority like the strategy board games. Despite the small subject pool, this background data indicates a broad range of preferences for school-based subject learning and out of school game play, and as such represents a decent sample for my experiments.

The last piece of background data concerns subjects’ self-reported attitudes on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree) in reference to 8 statements on problem and puzzle solving. According to Table 1, on average the subjects see themselves as oriented towards game playing and problem solving. They also see themselves as good observers and good with spatial relations.

Table 1: Puzzle and Problem Solving Self-Report

|Statement |Mean |Mode |

|I like crossword puzzles |3.6 |4 |

|I consider myself to be creative |4.1 |4 |

|I consider myself to be a good problem solver |4.2 |4 |

|I like to play board games |3.8 |4 |

|I am a good “on-your-feet” thinker |4 |4 |

|I am a keen observer |3.6 |4 |

|I am good with spatial relations |3.8 |4 |

|I like to play computer games |3.5 |4 |

Experiment 1 – The Training Study

Method

Participants

All 10 subjects participated in this study. They were divided into 2 conditions (6 training, 4 control) based on random assignment. Four of the subjects had played the game Set before, and 3 of the subjects had previously taken the Stroop Color/Word Test. None of the subjects had ever taken the Remote Associates Test.

Materials

An online version of the Stroop Color/Word Test (see Appendix B), a modified 15 question Remote Associates Test (see Appendix C) and the Set daily puzzle were used in the pre-test. The Set daily puzzle and the Set game cards were used in the training portion. The post-test consisted of 3 Stroop tasks (see Appendix B), a modified 15 question Remote Associates Test (see Appendix D), and the Set daily puzzle.

Design

A within-subject design was used, as subjects were observed prior to any treatment (pre-test), then they received a treatment, and finally they were observed at post-test. The independent variable was the treatment of training on Set game play or not, and the dependent variable was improved outcomes in Stroop, Remote Associates, or Set game play.

Procedure

The subjects were all given the pre-test during the week of 11/15/2004 – 11/21/2004. Starting on 11/21/2004, the training treatment group began playing the online Set puzzle every day and also participated in a few hours of Set card game play. They were to record their times on each daily puzzle. The last training puzzle was played on 12/10/2004, which was approximately 3 weeks from the start of training. On this day both groups were given a post-test.

Results

All Stroop tasks had two trials. The first was a “practice” trial where the words and colors (and later directions and animals) matched, and a second trial where the Stroop effect was brought into play. The control group averaged 21.2 and 25.6 seconds and the training group averaged 21.5 and 26 seconds on the 2 trials of the Stroop Color/Word pre-test task. The average time difference between the 2 trial runs was 4.4 seconds for the control group and 4.5 seconds for the training group. The control group averaged 8.25 and the training group averaged 9.5 correct responses on the 15 question Remote Associates pre-test (see Appendix C). The control group averaged 7:30 and the training group averaged 6:25 on the Set daily puzzle pre-test. Three subjects did not complete the puzzle after ten minutes, so their time was recorded as 10 minutes.

While the training group was asked to record their times during the daily puzzle task, the data is fairly incomplete as many subjects forgot at times, and a computer glitch caused the loss of one subject’s entire time record. What remains of the daily training data is presented in Appendix F.

The control group averaged 15 and 21.2 seconds and the training group averaged 15 and 23 seconds on the 2 trials of the Stroop Color/Word post-test task. The average time difference between the 2 trial runs was 6.2 seconds for the control group and 8 seconds for the training group. The control group averaged 10 and 13.6 seconds and the training group averaged 10 and 12.4 seconds on the directional Stroop task (see Appendix B). The control group averaged 13 and 19.3 seconds and the training group averaged 13.9 and 17.7 seconds on the animal Stroop task (see Appendix B). The control group averaged 10.75 and the training group averaged 12.2 correct responses on the 15 question Remote Associates post-test (see Appendix D). The control group averaged 8:00 and the training group averaged 1:56 on the Set daily puzzle post-test. One subject did not complete the puzzle after ten minutes, so his time was recorded as 10 minutes. A summary of the pre and post-test results is available in Appendix F.

Discussion

While the training group had a slightly smaller difference in time between the trials of the Stroop Directional and Stroop Animal tasks, the control group had a smaller difference in time between the trials of the Stroop Color/Word task. These differences are small enough that they could likely be explained by measuring inaccuracies inherent in the Stroop tasks used. In future trials, a more precise method of task administration would be highly desirable.

Both groups showed slightly improved performance on the post-test of the Remote Associates task. The improvement could be due to the fact that there were 2 repeated word triads (both groups improved, on average, by between 2 and 3 correct answers) though that would not explain the rest of the improvement. Another possibility is that all subjects improved because simply taking the pre-test helped them to understand the idea of convergent thinking. While the control group may not have been training on Set game play, they may have been working though the ideas behind the Remote Associates task during the experimental period. Finally, as this is a particularly small subject group and the pre and post-tests took place several weeks apart, a host of other confounding variables, such as what the subjects were doing in school or at work, could have influenced their convergent thinking in the interim. The most likely scenario is all subjects improved slightly because they understood the task better for the post-test than they had for the pre-test.

There is a notable disparity between the training group and the control group in Set game play of the daily puzzle. When the study began, the training group completed the pre-test puzzle on average a minute faster than the control group. There was over a 6-minute difference in average post-test puzzle times. This indicates, as hypothesized, that training on Set game play does improve performance on the daily puzzle.

While the results of this experiment are largely inconclusive, there are several aspects that are worthy of further study. Firstly, a larger sample size with more diverse attitudes towards problem solving and puzzle games would help counter the overwhelmingly positive enthusiasm these subjects reported. It is possible that an orientation towards puzzles invoked a ceiling effect in the training. Secondly, more accurate versions of the Stroop tasks and a broader selection of pre and post-test tasks would allow for further exploration of other possible mechanisms for Set game play. Thirdly, more training time or possibly more intense training might contribute to a more significant difference between groups on some of the tasks. Finally, it is possible that because the training task used the same Set cards with the same 4 attributes, the subjects were merely improving at pattern perception. In order to ensure true practice of executive functions, the attributes of the cards might need to vary across training trials.

Experiment 2 – The Think-aloud Study

Method

Participants

Two subjects participated in this study. They were both in the training group. One was an expert Set player who has been playing the game, though not on a regular basis, for nearly 8 years. The other participant was a novice set player who had been training on Set for 2 weeks when this experiment was conducted.

Materials, Design and Procedure

The standard 81 card Set deck was used for think-aloud game play. The experiments consisted of 2 trials of a modified version of a two person Set game. In both cases I was the second player. The subjects were to discuss each move before they made it, describe their mental model of the Set game board, and try as best they could to verbalize their strategies. Notes were taken during game play and in an immediate follow-up discussion on strategies.

Results

The Novice Player

The novice Set player (NP) approached the game with a very specific algorithm. NP would methodically search the board for sets of all ones, all twos, and all threes. He would look for these types of sets by counting the number of each type of card. If there were only 2 “ones”, or 2 red “ones “and” 2 green “ones,” he would rule out the presence of a “ones” set. NP then continued this process with the “twos” and “threes.”

When no sets of this type were found, NP would start with whatever “number” cards were least represented on the board. Moving from left to right he’d select a card from the least represented group (in this example a “one”), and then select the nearest “two”. Since, mathematically speaking, any 2 cards plus the appropriate third card define a set, he would look for the card that would complete the particular set. So if he had a “one-squiggle-shaded-green” and a “two-oval-open-red” he’d search for a “three-diamond-solid-purple.” If that card was not on the board, he’d start with the same one, pair it with a new two, and look for the third that would match. Once a set was found, three additional cards were placed on the board and the process would start all over again.

It is interesting to note that throughout game play, NP primarily attended to the number attribute of the Set cards. He would then examine, in shifting order, shape and color. The last attribute he would consider was the pattern. This attribute was greatly dwarfed by the other 3, to the point where all false sets NP made contained pattern violations. Another point of interest is that in the beginning of the game, NP completely reset his mental model of the board every time 3 new cards were put into play. For example, if a set he had just taken had no “ones” in it, and the 3 cards that replaced that set had no “ones” among them, he would still start over and count how many “ones” were on the board after every deal. As game play went on, however, he began to attend to this fact. During one of the final deals NP remarked to me “there are no new ones on the board, so I don’t need to look at that again.”

The Expert Player

The expert Set player (EP) approached the game with a variety of strategies. When the cards were first dealt, he took a moment to size up the board, and in the process picked up 2 sets having said almost nothing. When questioned as to how he found the sets, he told me that the ones he instantly sees are often in close proximity to each other. Sure enough, both sets he picked out were clustered right next to each other on the game board.

When EP did not see a set right away he would analyze the board to rule out what kinds of sets were not possible. He’d look at the attributes separately but in rapid succession, such as “there’s not a lot of green on this board, but there are a lot of solids and squiggles.” Each statement was referring to the same group of cards but across their various attributes. For EP, while he would name the attributes separately, the cards existed as all of their attributes simultaneously. Once EP had identified there were “no reds” or “only 1 squiggle,” he would no longer look for a whole class of set.

In regard to mental models of the Set game board, EP always kept the current board state in mind while looking at the new cards I dealt. Sometimes he’d be looking for a specific card he knew would complete a set, especially if the other 2 cards were currently in close proximity to each other. He did this twice during game play – called “set” before I had finished dealing the cards. When I asked about this, he told me he knew which cards he was looking for and as soon as I dealt them he had taken them.

Discussion

There are a few key differences in the novice and expert approaches to Set strategy even within this limited subject pool. Generalizing from this very small study, novices tend to search exhaustively for all possible sets, while experts narrow in on the types of sets they can make on any particular game board. Novices tend to see the cards as an ordered succession of attributes while experts view the cards as composed of concurrent attributes. Finally, novices tend to reconstruct their mental model of the board with each new deal, while experts merely update their mental model with the most current information.

The use of working memory is heavily implicated in the updating of the mental model shown by EP. Towards the end of the game, NP began to engage in this as well, though on a single-attribute basis. In either case, when a player has shifted to this type of consistent representation, Set card game play might be useful in training working memory. The representation of the attributes as either successive or concurrent might also have some implications for task-switching training, but this study did not deal with shifting in great depth. Further studies are warranted to both explore the role of other executive functions in Set game play and to gather more information on the variety of strategies used, including how a complete novice might approach the game.

Summary and Conclusions

There is a strong belief among parents and teachers that playing the game of Set helps to improve problem solving ability and strengthen critical thinking skills. This is the first study that attempts to empirically address the cognitive aspects of Set game play. While the results of the training study are mostly inconclusive, several ideas for further study have already been discussed. These include a broader subject pool, more accurate task measures, different training regimens, and different attributes on training cards. In addition, studying the near and far transfer possibilities of Set skills has great potential in linking Set to more general “real world” applications. Finally, while the think-aloud study provided a good basis for establishing novice and expert strategies in Set game play, some work remains in this arena as well, particularly in the realm of the complete novice. These ideas could all lead to a better understanding of the executive functions involved in Set game play and a more meaningful attempt at further training studies, an area of the field which has great potential. Targeting the executive functions Set game play makes use of and crafting a more nuanced training regimen is certainly an exciting and promising next step.

References

Bransford, J. D., Brown, A. L. & Cocking, R. R. (Eds.) (2000). How people learn: Brain, mind, experience, and school. Washington, D. C.: National Research Council.

Dempster, F. N., & Corkill, A. J. (1999). Interference and inhibition in cognition and behavior: Unifying themes for educational psychology. Educational Psychology Review, 11(1), 1-88.

Engle, R. W., Kane, M. J., & Tuholski, S. W. (1999). Individual differences in working memory capacity and what they tell us about controlled attention, general fluid intelligence and functions of the prefrontal cortex. In A. Miyake & P. Shah (Eds.), Models of working memory: Mechanisms of active maintenance and executive control (102-134). New York: Cambridge University Press.

Klingberg, T., Forssberg, H., & Westerberg, H. (2002) Training of Working Memory in Children with ADHD [Electronic version]. J Clinical and Experimental Neuropsychology, 24(6), 781-91.

Leahy, T. H. & Harris, R. J. (2001). Learning and Cognition, 5th edition. New York: Prentice Hall.

Mednick, S. (1962). The associative basis of the creative process. Psychological Review, 69(3), 220-232.

Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., Howerter, A., & Wager, T. D. (2000). The unity and diversity of executive functions and their contributions to complex frontal lobe tasks: A latent variable analysis [Electronic version]. Cognitive Psychology, 41, 49-100.

Olesen, P., Westerberg, H., & Klingberg, T. (2004). Increased prefrontal and parietal activity after training of working memory [Electronic version]. Nature Neuroscience, 7(1), 75-79.

Quinn, A. L., Weening, F., & Koca Jr, R. M. (1999). Developing Mathematical Reasoning Using Attribute Games. Retrieved on December 10, 2004, from

Van Brink, D. (1997) The Search for Set. Retrieved on December 10, 2004, from

Appendix A – Participant Questionnaire – November 2004

1) Age: Sex: 2) Level of schooling completed:

3) Favorite (up to 3) subject(s) in school? List subject and grade

level, aka high school or college:

4) Were you ever in an enrichment or gifted and talented

section/program in elementary school?

4a) If so, was work with puzzles or logic problems involved?

4b) If not, were puzzles or logic problems part of your elementary

school curriculum?

5) Favorite (up to 3) board games:

6) Favorite (up to 3) computer games:

7) Have you ever played "Set" before?

On a scale of 1 to 5, where:

1----------2----------3----------4----------5

Strongly Strongly

Disagree Disagree Neutral Agree Agree

For each statement, fill in the blank space with the number that

represents the degree to which you agree or disagree with the statement.

a) I like to do crossword puzzles:

b) I consider myself to be creative:

c) I consider myself to be a good problem solver:

d) I like to play board games:

e) I am a good "on-your-feet" thinker:

f) I am a keen observer:

g) I am good with spatial relations:

h) I like to play computer games:

Appendix B – Stroop Tasks

Stroop Color/Word:

Directional Stroop:

Animal Stroop:

Appendix C – Remote Associates Pre-test

Modified from:

01) Falling Actor Dust [ star ]

02) Coin Quick Spoon [ silver ]

03) Cracker Union Rabbit [ jack ]

04) Manners Round Tennis [ table ]

05) Salt Deep Foam [ sea ]

06) Ache Hunter Cabbage [ head ]

07) Barrel Root Belly [ beer ]

08) Color Numbers Oil [ paint ]

09) Ink Herring Neck [ red ]

10) Measure Desk Scotch [ tape ]

11) Board Magic Death [ black or game ]

12) Lapse Vivid Elephant [ memory ]

13) Magic Plush Floor [ carpet ]

14) Envy Golf Bean [ green ]

15) Sore Shoulder Sweat [ cold ]

Appendix D – Remote Associates Post-test

Modified from: and with thanks to Marie for one of them.

01) Falling Actor Dust [ star ]

02) Blue Cottage Mouse [ cheese ]

03) Widow Bite Monkey [ spider ]

04) Rabbit Cloud House [ white ]

05) Lick Sprinkle Mines [ salt ]

06) Coin Quick Spoon [ silver ]

07) Sandwich Golf Soda [ club ]

08) Speak Money Street [ easy ]

09) Hall Car Swimming [ pool ]

10) Cotton Bathtub Tonic [ gin ]

11) Diamond Cap “Hot Dog” [ baseball ]

12) Note Dive Chair [ high ]

13) Shopping Washer Picture [ window ]

14) Surprise Wrap Care [ package or present ]

15) Chocolate Fortune Tin [ cookie ]

Appendix E – Set Game Resources

Set Game Homepage:

Set Game Daily Puzzle:

Appendix F – Training Study Results

-----------------------

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

[pic]

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download