HED 489 Biostatistics

1
HED 489 Biostatistics

• Types of Data

First, make sure the audiois downloaded.
Then, start the audio and then start the PPT.
2
Types of Data

• Research usually consists of some form of
  measurement. Before one can understand effective
  evaluation, he/she must be aware of the different
  scales that measurement consists of.
• Nominal
• Ordinal
• Interval
• Ratio

3
Characteristics of Data
4
Nominal

• A number used for identification purposes only
  and has no meaning regarding amount of
  quality--attaches a label (doesnt imply value)
• Babe Ruth wore number 3 Barry Bonds wears
  number 25that doesnt make Barry Bonds 8 times
  betterits just an identification mark

5
Common Nominal Data in Public Health

• Gender
• 1Males 2Female
• Race
• 1African-American 2Caucasian
• Religion
• 1Jewish 2Christian 3Islamic

6
Ordinal

• Can do the same as Nominal AND
• attaches an order
• list of favorite teachers
• Top five foods
• Doesnt determine quantity of difference

7
Common Ordinal Data in Public Health

• Top Ten Causes of Death in the US
• All we know is that number 1 is the most frequent
  cause of deathwe cant tell the magnitude of
  such difference
• Top 5 Most Commonly Prescribed Medicines
• Birth Order

8
Interval

• Includes nominal and ordinal AND
• Possesses order with equality between points
• temperature--67 degrees is one degree warmer than
  66, as 34 is one degree warmer than 33
• IQ
• Many aptitude and attitude scales are interval
  based.

9
Ratio

• Almost identical to interval BUT
• order, with a zero.
• Income
• Age
• visits to clinics
• number of calories consumed
• number of cigarettes smoked
• lost pounds

10
Interval/Ratio

• Because of the similarities, and often confusion
  that exists, lets merge these two together

11
Test

• Sex
• Race
• Job
• Income
• Age
• Birth Order
• LOC
• Wellness Assessment
• Intelligence
• GPA
• Sperm Count
• Among the Most Likely to Succeed
• Preferred Top Three Birth Control Methods

Click to get answers
12
Test Answers

• Sexnominal
• Racenominal
• Jobnominal
• Incomeinterval/ratio
• Ageinterval/ratio
• Birth Order--ordinal
• LOCinterval/ratio
• Wellness Assessmentinterval/ratio
• Intelligenceinterval/ratio
• GPAinterval/ratio
• Sperm Countinterval/ratio
• Among the Most Likely to Succeedordinal
• Preferred Top Three Birth Control Methodsordinal
  

13
Converting Types of Data

• One can take this interval/ratio data and
  categorize it into one of four groups
• 0-18
• 19-30
• 31-50
• 51 higher
• You would then take your data and put it into one
  of these four groups, thus taking interval/ratio
  data and converting it to ordinal data

This is sometimes referred to as collapsing data
14

• The following are ordinal data
• 0-18
• 19-30
• 31-50
• 51 higher
• Its ordinal in the sense that you know that
  people in the groups below are older than the
  people above, but you dont really have a real
  understanding of how much older or younger. For
  example, somebody who is 31 years, 1 day is
  grouped in with a person who is 50 years, 364
  days.

15
Other common conversions

• Income
• Take their income and categorize into low,
  medium, or high socioeconomic status
• Weight
• Take their actual weight and categorize into
  normal, overweight, and obese

16
Why is this important to know

• The analysis of data will depend upon the type of
  data. More on this later, but to give you an
  example, lets say that we have 10 subjects, and
  weve labeled them as such

17
For sake of argument, weve labeled males with a
1 and females with a 2 All you are able to
do with this data is to report frequenciesyou
cannotreport the mean score of genderit would
make no sensethat mean would be 1.4 (14/10)
18
Data Analysis

• Generally speaking, this is the type of analysis
  you can do with the following data
• Nominal/Ordinal
• Frequencies chi-square
• Interval/Ratio
• Mean, standard deviation, t, F

19
Assignment 6

• Click here to pull up the sixth assignment.
  After completing the assignment, send it to me at
  kitt_at_siu.edu