stats Flashcards
nominal scale
ordering things based on name or form only (not quantative)
ex. hair colour
ordinal scale
set of ordered categories, not quantative. Only provides information about difference.
ex. satisfaction rating (dislike, neutral, like) or gold, silver, bronze.
Intreval scale
quantative, obtain all properties of ordinal measures. Zero does not represent absence of value
ex. tempterature (can go below zero).
Ratio scale
quantative, has an absolute zero (cannot go into the negatives)
ex. blood pressure/speed.
normal distribution
-symmetrical distribution of data in a bell shaped curve
-frequency of scores declines at extreme ends
-most subjects score in the middle
-mean, median & mode are essentially the same
-if data is normal 68% of the score will fall under the SD
mean
the average of a distribution of scores. (all scores added up divided by the amount of scores). sensitive to outliers.
median
the middle number in a set of scores. If there is an even number with two middle numbers the median will be the two middle numbers divided by 2.
mode
the most frequent number in a set of scores.
bimodal
when data has more than one mode
central tendancy
statistical measure that identifies a single value as representative of an entire distribution (inculdes mean, median & mode)
Skewed data
the mean no longer represents the centre of the distribution. The median or mode is a better reflection of the centre.
what are the measures of variability (2)
- range
- standard deviation
define range
-implies the difference between the highest and lowest score.
-influenced by outliers
-represented at highest number (minus) lowest number or highest number-lowest number
define standard deviation
provides information on the amount of variation from the mean.
ex. if mean is 10 and SD is 2 then the majority of the data in the set is around 8 to 12 (10-2=8, 10+2= 12).
1.0 SD is what percent of values?
1.96 SD is what percent of values?
2.58 SD is what percent of values?
1.0 SD= 68.2%
1.96 SD= 95.7%
2.58 SD= 99.6%
what does a larger SD mean?
larger SD means a larger difference in data/greater variability
define correlation (r)
-numerical co-efficient that indicates the extent to which variables are related
-indicates amount of relationship but not cause of relationship.
Always between -1.00 & +1.00
types of correlation (r)
a) +1.00
b) -1.00
c) 0.00
d) 0.5-0.7
e) 0.7-0.8
f) 0.8 and higher
a) positive correlation
b) negative correlation
c) no correlation
d) low correlation
e) moderate correlation
f) strong correlation
define and example of positive correlation (1.0)
-variables both increase or decrease together
(ex.) smoking and cancer risk (increase smoking and there is an increase cancer risk) &
Define negative correlation (-1.0)
when one variable increases the other decreases and vice versa
(ex.) Exercise and health risks. Increase of exercise decreases potential health risks
define validity
the accuracy of a measure or how well the results measure what they are supposed to measure
define reliability
the consistency of a measure or how repeatable are the results
can a test be valid if not reliable?
no, test must be reliable to be considered valid
can a test be reliable but not valid
yes, test can be repeatable but it doesn’t mean it’s measuring the right thing (ex) test V02MAX of runners but test them on bikes.
3 things to consider when considering a fitness test
- validity
- reliability
- normative data
Validity coefficient
-measures relationship between the scores from the test being validated to the criterion scores.
- r should be >8.0 to be considered valid
reliability coefficient
-relationship between trial 1, 2 & 3.
- r should be >0.90 to be considered reliable
why are normative standards important
-allows comparisons
-can see improvement relative to a group
-determines a rating for the fitness component
main sources of measurement error (4)
- test instrument…calibrated?
- test administrator…following protocols?
- subject…following test protocols?
- test environment…same warm up/time of day?