(Note: Please don’t give me anecdotes or personal experiences. These are the opposite of constructing a scientific test.)
To be tested: Whether music helps or hurts concentration for simple tasks.
I would like to make the test quantitative, so that it can be easily repeated on groups of school children or adults (I work in a library, and we have streams of potential volunteers.)
I thought of this one, but have been told it’s not really fair or accurate. Maybe I could modify it.:
Timed Test: pinpoint with highlighter on a map a set number of street intersections, chosen from a hatful at random.
Then add music and see if you can beat the time.
The problem is how to make it blind-blind, and I’m not real clear on that, having only seen the term described for tests on prescription drugs. I guess you would have to avoid stating what the test is about, maybe make it seem like a game, or that they were volunteering for real work.
Or please suggest a completely different test.
Thanks for your help.
I want to conduct it soon and have the results be meaningful.
Street intersections may vary significantly in the degree of ease with which they’re found, due both to the layout of the city and the test subject’s familiarity with the area. And, of course, testing a group of subjects in both conditions would require randomizing the testing sequence for each individual to help balance for practice effects.
A better task, that would negate differences between subjects’ familiarity with the mapped area, might be to flash a dozen 3-letter combinations on a screen and have them look for a specific set of say five and report how many out of the target set they saw. Changing the target set for each individual’s exposure to the different test conditions would help to control for practice effects, although I would still randomize the testing sequence and have multiple trials in each condition.
I can see that if you add “Elevator Muzac” then maybe somebody would simply ignore it like any other background noise, but make it something that sounds familiar (say “Jingle Bells”) which everbody will recognise and it has to become a different set of possibilities. Also do you class Alice Cooper or The Sex Pistols as music in the same way as Simon and Garfunkel or Johann Strauss.
I hate to admit that it makes a difference but for example I prefer to be in a shop that plays original recordings to one that plays tacky cover versions, even though the primary aim is to get the shopping… That would seem to imply that some people may find some types of music irritating or distracting but for reasons that are not necessarily due to the music.
I am often curious as to how some people seem able to concentrate on driving with 5000 watts of stereo so I’d think proper reserch could be interesting.
In order to make it double blind, you as the experimenter need to be unaware before hand whether a subject has music or not. So, they need to be in a room somewhere with their task and whatever measuring device to measure their productivity you come up with, and a computer to either play or not play music, but it doesn’t decide or tell you until you shut them in the room and go away. Perhaps the task could be some kind of computer game/work, for example
playing solitaire or some kind of simple video game.
Where’s cher3 - she should be able to give you a complete answer. In the meantime, thinking hard about group psych testing…
You’ll need a slide projector with a timer and paper and pencil.
Develop a series of word associations (pear + book. ruler + lamp). Present each pair to your subjects for a given period of time (say 4 seconds per pair). Go through the list once. At the end, give a piece of paper and a pencil to the subject and present the list again, using only one word from each pair (e.g., book or ruler). Subject must provide the associated word in a given amount of time (say, 8 seconds).
When testing, randomly assign each subject to hear music or not.
Do not tell the person interacting with the subjwcts what the goal of the experiment is (they can be out of the room when testing begins). Have the experimenter count the results (this is the second “blind” of a double-blind experiment - and the subjects are NOT told what the experiment is about until after the experiment is done).
Take the results and perform your stat analysis (a simple t-test for the most basic of setups). You can get more complicated by nesting your subjects according to classification (sex, major, etc). This will require a lot more subjects, but can also give interesting insight.
Report back here on your findings. This experiment is designed with the idea that presentation is standardized and the results are easily tabulated. Variations can be incorporated easily (ANOVA, using more than one kind of music or at different levels of loudness).
Whatever approach you take, do report back! And have fun!
I’d have the test takers take an IQ test. Make that a thorough IQ test that breaks intelligence into categories like verbal, analytical, spatial, etc. This way you can determine which tasks are hindered or helped by music. (i.e. Mozart might help with math, but not reading) Also follow the above advice about having a computer randomly turn music on for half the test takers when the test begins. With enough test takers, you should be able to see if music has any effect at all. Also, make sure it is the same music at the same volume every time.
Lance, the OP specified that, for this experiment at least, she’s specifically interesting in the effect of music on “concentration”, whatever that might mean. The other tasks might produce informative results, but not necessarily what she’s looking for. Depending on exactly what you’re trying to measure here, Fanny, I would recommend either an attention/memory test like the word assosciations proposed by brachyrhynchos, or time trials on some routine but tedious task, such as perhaps sorting a set of flash cards alphabetically-- Judging from your intersections example, I’d say that this is closer to what you’re looking for. I’d also recommend dividing your sample group into four subgroups: One does two trials without music, one does two with music, one does without and then with, and one does with and then without-- This should enable you, to some degree, to cancel out effects due both to individual difference and practice effects.
If you’re planning on trying it with different sorts of music, I would recommend that in addition to the fixed categories that you choose, you also have a category for “style of music enjoyed by subject”, and one for “style of music disliked by subject”, so, for instance, the results might be something like, people who listened to music that they liked showed a 30% decrease in times, while those listening to music they disliked showed a 10% increase.