Descriptive terms
Ex of random sample
Getting a list of all the high school music students in a major metropolitan area and selecting the names of 300 out of a hat or numbering the list and using a computer to generate 300 numbers/names to select.
Ex of stratified random
Getting a list of all the high school music students in a metropolitan area, dividing it into orchestra, band, and choir, then randomly selecting the names of 100 from each category
Haphazard
Group selected where all members of the population do not have an equal chance of being included in the research
Data
Information. In the research world, data, a plural term, are divided into four types: nominal, ordinal, interval, and ratio.
Standardized, Normalized Distribution -
Often referred to as the "bell curve," this set of results features an identical number of scores falling on each side of the mean with a preset number slightly above and below average (34%), a smaller group well above average and below average (14%), and a very few achieving outstanding or horrible results (2%). In this distribution the mean, mode, and median are identical. (see illustration on pg. 189).
Descriptive Research
Describes a population. Descriptive studies simply present characteristics of a sample with no attempt to show causal relationships among variables
Validity
Indicates whether an instrument or device measures what it is supposed to measure.
Statistics
Numbers used to summarize or provide a quick overview of a sample.
Branching
A survey tool that allows (or forces) respondents to skip ahead to questions that apply to them
Standard Deviation
A measure of variability. A number used to tell how measurements for a group are spread out from the mean. A low standard deviation means that most of the numbers are close to the average. A high standard deviation means that the numbers are more spread out.
Rating Scal
A research tool that allows the researcher to measure the opinions and behaviors of respondents in a quantitative manner.
Closed Questionnaire
A research tool that may ask respondents to answer "yes or no", circle or otherwise identify an answer, rank items according to their correctness or validity, or insert specific data into a blank.
Random
All members have an equal chance of being included in the research.
Reliability
An indication of whether an instrument or device will show the same results under identical or similar conditions.
Ex of skewed data
Because the temperature of the sight reading room was 58 degrees (the heater broke the night before the contest) the scores may have been skewed toward a lower than expected level of performance. Or, because the person conducting the one-on-one survey regarding music's popularity in a given middle school was a guest at the school who happened to be the lead singer for Bon Jovi, the results may have been skewed.
Biased Sample
Biased: Intentional or unintentional overselection of subjects from a given strata
Ex: of interval/ratio data
Counting the number of music terms students can correctly match with given definitions on a test.
Interval/Ratio
Data which can be compared to determine order and which provide enough information to calculate the exact distances between each data point. The only difference between interval and ratio data categories is that the first has an arbitrary starting point and the second has an absolute zero.
Ordinal
Data which can be compared to determine rank order (faster, smarter, better), but the degree to which one is better, faster, smarter, etc. cannot be determined.
Nominal
Data which can be given a name or category designation but which typically are not compared regarding better, faster, smarter, louder, etc.
Internal Validity
Do changes in the independent variable account for changes in the dependent variable? Example: In measuring graduate versus undergraduate voice students' decibel levels while singing forte you find that both groups produce amazingly similar results of 85-86 decibels. Then you find out that your decibel meter is broken and all sounds, despite their strength register between 85-86d decibels. Your data lack internal validity.
Ex: of nominal data
Example: Counting string players versus wind players versus percussionists versus combination in beginning 5th graders' applied instrument choices
External validity
External: Do research techniques used and the sample employed reflect the real world? Example: You complete a study using Gordon's rhythmic solfege syllables and find that students employing them sight read significantly better than students using Kodály rhythmic solfege syllables. However, your colleagues show little interest since it took 4 months for the students to learn all the rhythmic permutations in the Gordon model. They are implying that your study has little external validity (i.e., it is not practical). They are even more convinced not to attempt your approach when they discover that you used a biased sample including only students with IQs over 130.
Ex of haphazard sample
Surveying a group of students at the mall regarding how they feel about K-12 music programs and inferring it to the general population.
Ex of biased sample
Surveying a list of all-state music students regarding how they feel about high school music programs and inferring the results to the general population of high school music students
Sample Size
The number of subjects in your study (e.g. N = 100 students)
Skewed Data
Unlike the Standardized, Normalized Distribution, skewed results mean that the results are not symmetrical. Typically skewed results mean that a large proportion or even a majority do very well or very poorly on a given task. Sometimes the term skewed is used interchangeably with the word "biased" meaning that some information or activity caused the results to appear higher or lower than they actually are.
Ex: of ordinal data
comparing the seating placements of students in your high school band from two feeder middle schools.
Stratified Random
the population is first divided into categories to ensure representation of all subgroups of interest to the research in the study.
