Unit 4 sec 2.5
Single and paired data
22 November 2015
14:56
Looking just at column H, the values are all based on a
single measure (weight) and can be described as single data, you could calculate an average: you will more about
averages in section 3, were 2 different types of average RX warned
alternatively you could measure how widely dispersed. The values are - in other
words, whether the values are tightly clustered together. Why waste bright.
Finally, you could plot the values to discern the overall pattern visually and
you will be shown a number of useful statistical plots in unit 11. The purpose
of doing these things would be to try and gain an insight into baby weights in
general.
Suppose now that there was a 2nd sample, birth
weight grumpy different set of mothers, the babies in the 2nd sample
of being classed as premature. This is now a two sample, as opposed to one sample dataset. Other examples of
making statistical comparisons might include making a comparison between 2
medical treatments or to commercial products. Again, a sample of measures would
be taken from each of the results compared. (When making such a statistical
comparison, there is no requirement that the 2 samples contain the same number
of values, although they could do.)
A statistical question of interest might be held the baby’s
weight (in column H) relate to the way of their mothers at the start of their
pregnancy (column F). This question links to pieces of information for each of
the people in this study - a case of
paired data. In statistical investigation terms, this falls under the
general heading of seeking a relationship. A number of important statistical
ideas are linked to exploring relationships today with paired data.
In particular, how to classify and distinguish different
types of data. Primary and Secondary data are terms that identify the data
source - primary data you collect
yourself, whereas secondary data are taken from somewhere else.
No comments:
Post a Comment