linxit (linxit) wrote in times_xwd_times,

Crossword Research update

I heard from Kathryn Friedlander this morning with an update to the research on crossword solving skills that some of us took part in a couple of years ago. For those who didn't, this involved solving a couple of general problems then solving a crossword against the clock while being filmed, and having to talk through the thought processes while solving.

Here's a copy of a press release following initial presentation of the results at a couple of conferences. The full paper will be made available in due course.

Problem-solving ability is key to Cryptic Crossword Success

Dr. Philip Fine and Dr. Kathryn Friedlander, University of Buckingham


Fine, P. & Friedlander, K. J. (2013, 28-31 August). Cryptic Crossword Expertise and Fluid Intelligence. presented at the International Symposium on Performing Science, Vienna; proceedings paper published by the Royal College of Music, London (in print).

Friedlander, K. J. & Fine, P. (2013, 4th-6th September). Fluid Intelligence and Cryptic Crossword Expertise. presented at the CogDev 2013: BPS Developmental and Cognitive Sections Joint Conference 2013, Reading

Full Details of the Research

Our study falls into the broad category of ‘Expertise Research’. Many previous studies have looked at how people become ‘expert’ in practice-intensive fields such as music, chess or Scrabble, and have weighed up the relative contributions of hard work, motivation and talent. Based upon these studies, it has become commonplace in popular psychology (for example, Malcolm Gladwell’s ‘Outliers’) to downplay the importance of innate aptitudes and to stress the need for regular, focused practice routines over many hours (the 10 year/10,000 hour rule), starting from early childhood.

Cryptic crossword solving offers a fresh approach to the study of expertise because the ground-rules are rather different. Solvers of all standards generally begin tackling puzzles in their teens or early twenties, and spend only 5-10 hours per week solving puzzles; yet there is a wide range of expertise. Our previous surveys also indicated that cryptic crossword solvers were generally very well-educated and tended to favour university subjects and careers in the fields of IT, science and maths. This led us to hypothesize that cryptic crossword solvers as a whole were particularly good at logical problem-solving, compared to the normal population; and to suspect that expert solvers might be even more proficient at this.

We invited 28 participants to the University of Buckingham, comprising 10 non-expert solvers of cryptics, who generally take longer than 30 minutes to solve a standard daily cryptic crossword; and 18 experts, representing an elite body of top-class performers in the field. These elite solvers could be divided into the following 3 groups (although there was some overlap between them): 6 speed-solvers, who compete successfully at the Times Speed-solving Championships (including 3 past champions); 6 solvers of advanced cryptics such as the ‘Listener’ crossword (including a number of ‘annual all-correct’ and ‘roll-of-honour’ solvers); and 6 people who either set or edit crosswords professionally, as well as being successful solvers in their own right. We made sure that all participants had equivalent experience (both experts and non-experts had solved for an average of over 30 years) and that they spent similar amounts of time solving crosswords each week (about 7.5 hours on average).

We asked our participants to complete both parts of the AH5 test of problem solving (Heim, 1968), a timed high-grade fluid intelligence test which measures the ability to solve novel problems ‘on one’s feet’. The AH5 is intended to discriminate among very highly able people at university/post-graduate level. Part I measures verbal and numerical reasoning abilities, and Part II non-verbal (diagrammatic) reasoning abilities.

Then, after two warm-up word games, participants solved a bespoke cryptic crossword which we commissioned from a setter for the Independent newspaper. The puzzle was typical of those found in a good-quality daily broadsheet newspaper. The participants were allowed 45 minutes to solve it and were filmed while they did so, talking aloud to explain their thoughts and progress. We are still in the process of analysing these recordings, which will form the basis of a separate article.
Analysing the AH5 scores, we found that all the participants scored well within Heim’s range for high-ability populations, indicating that all participants were indeed excellent problem-solvers. But, as we hypothesized, expert solvers scored significantly higher overall (and in Part I, but not in Part II) than the non-experts. The experts’ scores placed them, as a group, in the very top band of Heim’s range.

Most of the non-expert solvers (8 out of 10) failed to finish the crossword in 45 minutes; conversely 17 out of the 18 experts finished. The finishing times ranged from 11:07 to 40:35 minutes. The solving time of the finishers also correlated significantly with their Part I AH5 scores: those who finished faster generally had higher AH5 Part I scores.

We have concluded from this that cryptic crossword solvers as a whole appear to be of above average fluid intelligence, and thus may have an innate aptitude for problem-solving which makes cryptic crossword solving an attractive and rewarding pastime. Experts score even higher on fluid intelligence tests.

As with all skills, learning to solve cryptic crosswords requires practice over a number of years to acquire a good knowledge of the rules and the common devices. But both of our groups had equivalent experience of solving cryptic crosswords over many decades, yet their performance in tackling our bespoke crossword was strikingly different. This suggests that experience per se does not fully explain group differences in this domain, and that fluid intelligence may play a significant role. We are now looking at some of the sub-skills which the AH5 may have been tapping into.

  • Post a new comment


    default userpic

    Your IP address will be recorded 

    When you submit the form an invisible reCAPTCHA check will be performed.
    You must follow the Privacy Policy and Google Terms of use.