— by sorceress

This survey was run for two weeks in September 2014, during the rating phase of LD30.

right-click -> view, to embiggen any image.


















How you got through the weekend



About your tools and methods



About your game



Satisfaction

Satisfaction scores are generally high. People on the whole are happy with the choices they made, and the work they did. Also see that people are mostly happy with the LD website, and with the theme choice.


Difficulties

People on the whole experienced few difficulties. The most common difficulties were experienced trying to come up with an idea, composing sounds/music, and wasting time on minor details.


Post-Mortem

The Post-Mortem is mostly optimistic. Ludum Dare is a learning experience for many. Most have identified areas they spent not enough time on, which they will hopefully take into consideration in future game jams.


Heat Maps

Here are some charts showing the correlation between pairs of responses. It would be impractical to show every pairing, so what is shown here are a selection of the strongest and most interesting.

The values computed are the product-moment correlation coefficients. The majority of pairs actually show very little correlation, with values typically below 0.1 in magnitude.

Gender shows no real correlations with anything. All correlation coefficients involving Gender are below 0.1 in magnitude.

Age -vs- Education: is included because it is such a nice chart.

Education -vs- Skill rating: shows a small positive correlation. On closer inspection, see that people with degrees are much more likely to consider themselves game-dev experts, than people without degrees.

Worked at night -vs- Got enough sleep: shows a moderate negative correlation as we might expect. Working at night can be tempting for various reasons, but you’re likely to feel significantly impaired for doing so.

Used familiar language -vs- Used familiar arting tools: shows a moderate positive correlation. The vast majority did use familiar language and arting tools, so perhaps the correlation comes from those completely new to ludum dare and game-dev? ie, First timers who not familiar with any language or arting tools.

Has sound -vs- Happy with sound: shows a fairly strong positive correlation. But perhaps this is not as interesting as it seems. Being happy with your sounds strongly suggests that you have sounds.

Happy with theme -vs- Had trouble coming up with idea: shows moderate negative correlation. Naturally, people are happier with a theme IF they can come up with ideas for it. If you’re stuck, blame the theme!

Finished game -vs- Happy with game: shows a moderate positive correlation. Finishing makes people happier. If you want your experience of LD to be as positive as possible, then make sure you take on an idea you can actually finish!


All correlations

Here we have a graph showing the correlation coefficient between all pairs of question responses. Red is a strong positive correlation. Blue is a strong negative correlations. White is no correlation.

It is interesting to note the ‘halo effects’ in the upper right corner. The responses to the ‘Satisfaction’ block of questions show a weak positive correlation with all others in that block. The same effect is seen in the ‘Difficulties’ and ‘Post-Mortem’ blocks.

We might conclude that a person feeling satisfied about one area of their game is more likely to feel satisfied about other areas of their game as well. Likewise, experiencing difficulties with one area of game-dev makes one likely to experience difficulties in other areas too.

Also see there is a negative correlation between the ‘Satisfaction’ and ‘Difficulties’ blocks, suggesting that it is harder to feel satisfied with any aspect of your game, if you experience difficulties in any area of making your game.


Correlations with Ratings

Many survey respondents supplied their username with the form, allowing me to find correlations between their survey responses and the ratings their LD30 game received after the voting stage had completed.

While interesting to explore, unfortunately the correlations were all rather weak, with these six being the strongest.

sol_1

sol_2

sol_3

sol_4

sol_5

sol_6


[cache: storing page]