A STATISTICS REFRESHER  A scale has the property of magnitude if

we can say that a particular instance of the

WHY WE NEED STATISTICS attribute represents more, less, or equal
 Statistics are important for purposes of amounts of the given quantity than does
education another instance
 Numbers provide convenient summaries EQUAL INTERVALS
and allow us to evaluate some  A scale has the property of equal intervals
observations relative to others if the difference between two points at any
 We use statistics to make inferences, place on the scale has the same meaning
which are logical deductions about events as the difference between two other points
that cannot be observed directly that differ by the same number of scale
 Detective work of gathering and displaying units
clues – exploratory data analysis  A psychological test rarely has the
 Then confirmatory data analysis property of equal intervals
 Descriptive statistics are methods used to  When a scale has the property of equal
provide a concise description of a intervals, the relationship between the
collection of quantitative information measurement units and some outcome
 Inferential statistics are methods used to can be described by a straight line or a
make inferences from observations of a linear equation in the form Y=a+bX
small group of people known as a sample  Shows that an increase in equal units on a
to a larger group of individuals known as a given scale reflects equal increases in the
population meaningful correlates of units
SCALES OF MEASUREMENT ABSOLUTE 0
 MEASUREMENT – act of assigning  An Absolute 0 is obtained when nothing of
numbers or symbols to characteristics of the property being measured exists
things according to rules. The rules serves  This is extremely difficult/impossible for
as a guideline for representing the many psychological qualities
magnitude. It always involves error. SCALES OF MEASUREMENT
 SCALE – set of numbers whose NOMINAL SCALE
properties model empirical properties of  Simplest form of measurement
the objects to which the numbers are  Classification or categorization
assigned.  Used for naming or describing things.
 CONTINUOUS SCALE – interval/ratio. A They simply classify subjects into
scale used to measure continuous categories.
variable. Always involves error  Arithmetic operations can be performed
 DISCRETE SCALE – nominal/ordinal with nominal data
used to measure a discrete variable (ex.  Ex.) Male or female
Female or male)  Also includes test items
 ERROR – collective influence of all of the  Ex.) yes/no responses
factors on a test score. ORDINAL SCALE
PROPERTIES OF SCALES  Classifies in some kind of ranking order
MAGNITUDE  Individuals compared to others and
 The property of “moreness” assigned a rank
 Imply nothing about how much greater one  Cannot be used to make statements about
ranking is than another ratios
 Numbers/ranks do not indicate units of DESCRIBING DATA
measure  Distribution: set of scores arrayed for
 No absolute zero point recording or study
 Binet: believed that data derived from  Raw Score: straightforward, unmodified
intelligence test are ordinal in nature accounting of performance, usually
INTERVAL SCALE numerical
 In addition to the features of nominal and FREQUENCY DISTRIBUTION
ordinal scales, contain equal intervals  Frequency Distribution: All scores listed
between numbers alongside the number of times each score
 Assign numbers to indicate whether occurred
individuals are less than, greater than, or  Grouped Frequency Distribution: test-
equal to each other. score intervals (class intervals), replace
 No absolute zero point the actual test scores
 E.g. A person’s level of anxiety,  Highest and lowest class intervals= upper
intelligence or of psychopathology and lower limits of distribution
 Can take average  A distribution of scores summarizes the
 Shows difference between scores scores for a group of individuals
RATIO SCALE  Frequency distribution: displays scores on
 In addition to all the properties of nominal, a variable or a measure to reflect how
ordinal, and interval measurement, ratio frequently each value was obtained
scale has true zero point  One defines all the possible scores and
 Equal intervals between numbers determines how many people obtained
 Ex.) measuring amount of pressure hand each of those scores
can exert  Histogram: graph with vertical lines drawn
 Have . ‘true’ value of zero. at the true limits of each test score (or
 E.g. income level, level of performance class interval) forming TOUCHING
 True zero doesn’t mean someone will rectangles- midpoint in center of bar
receive a score of 0, but means that 0 has  Bar Graph: rectangles DON’T touch
meaning  Frequency Polygon: data illustrated with
NOTE: PERMISSIBLE OPERATIONS continuous line connecting the points
 Level of measurement is important where test scores or class intervals meet
because it defines which mathematical frequencies
operations we can apply to numerical data  A single test score means more if one
 For nominal data, each observation can be relates it to other test scores
placed in only one mutually exclusive  Whenever you draw a frequency
category distribution or a frequency polygon, you
 Ordinal measurements can be must decide on the width of the class
manipulated using arithmetic interval
 With interval data, one can apply any MEASURES OF CENTRAL TENDENCY
arithmetic operation to the differences  Measure of central tendency: statistic that
between scores indicates the average or midmost score
 between the extreme scores in a  First quartile is the 25th percentile;
distribution. second quartile is the median, or 50th
 The Arithmetic Mean percentile; third quartile is the 75th
 “X bar” percentile
 sum of observations divided by number of  The interquartile range is bounded by the
observations range of scores that represents the middle
 Sigma (X/n) 50% of the distribution
 Used for interval or ratio data when  Deciles are similar but use points that
distributions are relatively normal mark 10% rather than 25% intervals
THE MEDIAN  Stanine system: converts any set of scores
 The middle score into a transformed scale, which ranges
 Used for ordinal, interval, and ratio data from 1 to 9
 Especially useful when few scores fall at  The Average Deviation AD= ∑ ▒|x|/n
extremes  X-mean=x
THE MODE  Average deviation= (sum of all deviation
 Most frequently-occurring score scores)/ total number of scores
 Bimodal distribution- 2 scores both have  Tells us on average how far scores are
highest frequency from the mean
 Only common with nominal data  Variance: the mean of the squares of the
MEASURES OF VARIABILITY difference b/w the scores in a distribution
 Variability: indication of how scores in a and their mean
distribution are scattered or dispersed 
 Measure of variability: any statistics that
describe the amount of variation in a
distribution 
 The Range
 Difference between the highest and lowest
scores  Found by squaring and summing all the
 Quick but gross description of the spread deviation scores and then dividing by the
of scores total number of scores
 The interquartile and semi-interquartile  s = sample standard deviation
range  σ = population standard deviation
 Distribution is split up by 3 quartiles, thus  The Standard Deviation
making 4 quarters each representing 25%  Similar to average deviation
of the scores  But in order to overcome the (+/-) problem,
 Q2= median each deviation is squared
 Interquartile range measure of variability  Standard deviation: a measure of
equal to the difference between Q3 and variability equal to the square root of the
Q1 average squared deviations about the
 Semi-interquartile range interquartile mean
range divided by 2  Is square root of variance
 Quartiles and Deciles
 Quartiles are points that divide the
frequency distribution into equal fourths
 where the latter has arbitrarily set mean
and standard deviation
 used for comparison
Z-SCORE
 conversion of a raw score into a number
indicating how many standard deviation
units the raw score is below or above the
mean of the distribution.
SKEWNESS  The difference between a particular raw
 skewness: nature and extent to which score and the mean divided by the
symmetry is absent standard deviation
 POSITIVELY SKEWED Ex.) test was too  Used to compare test scores with
hard difference scales
 NEGATIVELY SKEWED ex.) test was too T-SCORE
easy  Standard score system composed of a scale
 can be gauges by examining relative that ranges from 5 standard deviations below
distances of quartiles from the median the mean to 5 standard deviations above the
KURTOSIS mean
 steepness of distribution  No negatives
 platykurtic: relatively flat OTHER STANDARD SCORES
 leptokurtic: relatively peaked  SAT
 mesokurtic: somewhere in the middle  GRE
 Linear transformation: when a standard score
THE NORMAL CURVE
retains a direct numerical relationship to the
 Normal curve: bell-shaped, smooth,
original raw score
mathematically defined curve, highest at
 Nonlinear transformation: required when data
center; both sides taper as it approaches are not normally distributed, yet comparisons
the x-axis asymptotically with normal distributions need to be made
 -symmetrical, and thus have mean,
median, mode, is same
 Area under the Normal Curve
 Tails and body

 STANDARD SCORES
Standard Score: raw score that has been
converted from one scale to another scale,
THE NORMAL CURVE AND STANDARD SCORES
 Standard scores involve measurement on
an interval scale and their norms
represents scores having means and
standard deviation chosen for their
usefulness.
 Scales gained from a measure which can
be referenced to the normal distribution
and are said to be normalized.
 Norms are usually expressed in terms of
either z scores, T scores, sten and stanine NORMALIZING SCORES
scores or percentiles.  Normalized score provide a common
Z SCORES “Z VALUE standard scale for different tests. They are
 Most basic of all standard scores. distributed along an equal interval scale
 Standard curve has a mean of 0 and a and arithmetic operations can be carried
standard deviation of 1. out.
 It becomes easy to convert from any other  Because they are normalized, a particular
normal curve to the standard z curve, we z score always means the same thing in
subtract the mean and divide by SD. percentile terms.
 Z score are useful, showing how many
SD’s away from the mean a score is. A
positive value suggest that someone is
above average while a negative value is
below average.
T SCORES
 Provide a scale having a mean of 50 and a
standard deviation of 10.
 Having a range from about 20 to 80, within
 the range of three SDs either sides of the
 As Figure 4.13 shows, at three SD’s below mean.
 T scores often being rounded to the
the mean we have -3, at two SD’s below nearest whole numbers.
we have -2, at one below we have -1, at  T = 50 x test score – mean / SD
the centre of the curve we have 0, then at STEN SCORES
 Standard Ten “STEN” scores are widely
one SD above we have 1, at two above we
used, esp. in personality assessment
have 2 and at three above we have 3. because they provided the scale most
Values in between the whole numbers can preferred by Cattell in designing the
occur so that a person’s score could be at original 16PF.
 5.5 is the mean value of the scale,
1.55 or -2.14 or 0.57, for example. Based because they go from 1 to 10. It has an SD
on it’s definition, a z-score of 1.5 is 1.5 of 2.0 so a range of two SDs below the
above the mean. mean occurs at value of 1.5 and two
above at 9.5.
STANINE SCORES
 “standard of nine”, form a scale which
goes from 1 to 9.
 Dividing the normal curve and having an
SD of 2.0, the mean is set at 5 on the
scale.
 A range of two SDs below the mean
occurs at 1 on the scale and two above at
9.
CONVERTING RAW SCORES TO
STANDARD SCORES
 To convert a z score to a sten score:
 1. Multiply the z score by 2
 2. Add 5.5
 3. Round to the nearest whole number
 4. Values greater than 10 are given as 10 DATA ANALYSIS IS ONLY ONE PART OF
and value less than 1 are given as 1. THE RESEARCH PROCESS
 To convert a sten back to a z score: BASIC CONCEPTS OF STATISTICS
 1. Subtract 5.5 from the sten score DATA
 2. Divide the result by 2  Is observed characteristics of interest
 To convert a z score to T score:  For example, academic performance,
 1. Multiply the z score by 10 water quality, marital status
 2. Add 50 POPULATION
 3. Round to the nearest whole number  Refers to the total sets of subjects of
 4. Values greater than 100 are given as interest in a study
100 and values less than 1 are given as 1.  For example, all engineers in denr, all
 To convert a T score to z score: public school teachers in manila
 1. Subtract 50 from the T score SAMPLE
 2. Divide by 10  Is the subset of the population on which
 In areas of practice, esp. clinical field, it is the study collects data
important to have a good understanding of  For example, 20 out 100 engineers in
the properties of the normal curve and of denr, 60 out 1,000 public school teachers
standard scores so a to take care when in manila
interpreting extremely low or high SUBJECTS
assessment scores.  Refers to the entities of study observed
 There are other forms of non-normality  For example, college students, companies,
except for skewness such as multi- families
modality or a uniform or near-uniform. QUALITATIVE VARIABLES
 Variables that involves characteristics or
BASIC CONCEPTS ON STATISTICS attributes
 For example, sex (male/ female),
RELAVNCE OF STATISTICS geographic location (ncr, region ii, region
 Is consists of body of methods for v) and so on
obtaining and analyzing data QUANTITATIVE VARIABLES
 Helps in decision making to various  Variables are numerical and can ranked or
problems ordered
 Provide methods for design, description,  For example, age, heights, body
and inferences temperature, car speed
THE RESEARCH PROCESS IMPORTANT TERMS IN STATISTICS
INDEPENDENT VARIABLES
 A variable thought to be the cause of some
effect
DEPENDENT VARIABLES
 A variable thought to be affected by the
changes in
 An independent variable
LEVEL OF MEASUREMENT
 The relationship between what is being
measured and the numbers that represent
what is being measured
MEASUREMENT
 Assignment of numbers to objects or
events using rules to represent specified
attributes of the objects
ATTRIBUTES
 Dimension along which individuals can be NON PROBABILITY SAMPLING
measured and along which they vary  Is a sampling technique in which the
chance or probability of each unit to be
selected is not known or confirmed
THE RELATIONSHIP OF THE VALUES THAT
ARE ASSIGNED TO THE ATTRIBUTES FOR
A VARIABLE

MULTISTAGE SAMPLING
 Refers to a combination of two or more
sampling technique
CATEGORICAL VARIABLE
 Is one that names distinct entity
 Male or female
NOMINAL VARIABLE
 Is one that names distinct entity but with
more than
 Two categories
 Catholic, islam, born-again christian
ORDINAL VARIABLE
 The categories has logical order
 1st place, 2nd place, 3rd place
INTERVAL VARIABLE
 Equal intervals on the variable represent
equal differences in the property being
measured
 Weight: 45kg-50kg; 60kg-65kg
RATIO VARIABLE
 The same as an interval variable, but the
ratios of scores on the scale make sense
 Score on resilience scale of 40, in reality is
more happy on someone scoring 10
DATA COLLECTION METHODS
TELEPHONE SURVEY
ONLINE SURVEY
SAMPLING TECHNIQUES
PROBABILITY SAMPLING
 Is a sampling technique in which each unit
has equal chance of probability of being
selected